**6.3 Batteries**

Batteries are an important component of the AVIS in order to supply the power to the motor during off sunshine hours. Generally, lead-acid batteries and deep cycle batteries of types AC and DC are most frequently used in solar-powered irrigation systems [63]. The lead-acid battery is 80% efficient which reflects in storing 25% more energy in the battery. However, in most studies deep cycle batteries are recommended for solar energy storage due to their prominent attributes such as being discharged to a low energy level, rapid recharged, and no regular maintenance or topping-up required. The capacity of the battery should be sufficient enough to bear the load and smoothly run the connected appliances. If properly maintained, a deep

cycle battery can prolong up to 3–4 years compared to the lead-acid battery, which is limited to 2–4 months.

#### **6.4 Solar motors and pumps**

Solar-driven pumps either surface or submersible are the heart of the AVIS. The harvested solar energy is directly being utilized to drive the AC or DC motors, which mainly transform the electrical energy into mechanical energy. The mechanical energy is then utilized to develop hydraulic energy for lifting the groundwater from the deep levels. If compared submersible pumps lift groundwater from more depths as compared to the surface pumps. However, surface pumps or centrifugal pumps are more often adopted due to their special attributes. The selection of the pumping unit is mainly dependent upon the groundwater depth. Higher groundwater depth needs high power pumps and vice versa. Different vendors are available in the market such as Shurflo, Grundfos, Lorentz, Dankoff, SolarJack, etc., [63] which produce different capacities of the solar pumps.

#### **6.5 Sun tracker**

The earth revolves around its orbit thereby, the solar radiation incident on the solar collector changes. The sun tracking device is vital in this case since it aims to direct solar energy perpendicularly to the sun during the entire sunshine hours. The device is capable to improvised the solar energy extraction 10−70% [19]. The cited study reveals that the sun tracker generates 57% more, however, increases the installation and maintenance cost of the irrigation system [64]. The sun trackers are classified based on the single axis and dual axis differentiated based on the degree of freedom **Figure 12**. The sun trackers are coupled with photo sensors, which create voltage difference when solar radiation incident on it and accordingly adjust its best orientation. Dual axis sun tracker entails two axes, primary axes and a secondary axis. The primary axis adjusts the solar panel with respect to ground rotation whereas the secondary axis provides tilt movement of the solar panel. However, it has been equipped with installation complexities and also mounts the project cost.

#### **6.6 Micro controller**

In order to ensure the smooth operation of the solar pump, it is essential to have both a maximum power point tracking system (MPPT) and a variable frequency inverter (VFI). Various configurations of variable frequency drive (VFD) are being investigated by coupling with and without MPPT. The VFD controller provides square wave output, which results in higher-order harmonics in the output, causing extra losses and pulsating torque in the motor. On the other hand, the maximum power point was not able to identify if the MPPT is not mounted, consequently, the system operates at a fixed DC voltage. Furthermore, the groundwater head found an influential entity for manipulating the average power tracking efficiency. Yadav et al. [67] investigated and compared the impact of sine wave MPPT and VFD on the tracking efficiency corresponding to varying the water head. It was reported that the tracking efficiency ranges between 99.30 and 99.60% when the water head fluctuates from 10 to 20 m. However, in the case of VFD tracking efficiency dropped to 72.20%. In addition, the sine wave MPPT coupled with a low pass filter eliminates the higher-order harmonics.

**Figure 12.** *Single axes (a) and dual axes (b) sun tracker [65, 66].*

The MPPT was also found sensitive to environmental variables such as shading and temperature. In this context, various tracking algorithms are being investigated. The incremental conductance method (INC) works on the principle of comparing how voltage and current change. The INC technique, however, has a step size difficulty, particularly in rapidly changing weather conditions. Aside from that, the control system necessitates a costly and intricate circuit and is incapable of dealing with partially shadowed conditions, such as the shadow created by clouds and trees. Recently, it was realized that dual MPPT coupled with INC and dormant particle swarm optimization (DPSO) could handle the partial shading effect and mitigate the voltage fluctuation and spikes [68]. MPPT based on extremum-seeking control has a rapid convergence time and strong steady-state performance because the operating voltage or current of solar PV arrays can be dynamically tuned to maximize output power. Gray Wolf 's optimization-based MPPT method was investigated and found efficient in terms of fewer operating parameters, higher efficiency, and outperformed under partially shaded conditions [69]. The artificial neural network (ANN) based MPPT method was also investigated for MPPT [70]. The ANN obtained records of environmental variables from sensors and give the output signal of maximum power point generating conditions. Ramaprabha et al. [71] also investigated the ANN algorithm and found it efficient for all insolation levels. Lin et al. [72] proposed a radial basis function network (RBFN) and an improved Elman neural network (ENN) for MPPT and found them effective. Similarly, another study concludes the ENN implementation due to its stable response and low fluctuation as compared to perturb and observe (P&O). One can find the details relevant to various types of ANN models from the cited articles [73, 74]. However, the variable step size-based ANN MPPT possesses high accuracy, with a quick convergence response time as compared to the step sizebased ANN MPPT algorithm [73].

The microcontroller can also be used to determine the optimum irrigation requirements. Rajkumar et al. [75] investigated an automated switching mechanism of the water pump by collecting data from the temperature, humidity, and soil moisture sensors. It was reported that the microcontroller successfully regulates the pumping time. Gao et al. [76] designed a fuzzy irrigation control strategy for real-time estimation of real crop water requirements based on the data received from sensors. Based on experimental results it was realized that the system precisely regulates the solenoid valve, which turns to improve efficiency and proven food security. Normally, *Agrovoltaic and Smart Irrigation: Pakistan Perspective DOI: http://dx.doi.org/10.5772/intechopen.106973*

the sensor data contains noise that needs the preprocessing of the collected data. In this context, the recursive adaptive filter method was employed to remove the noise, handle the missing values, and normalize the data [77]. Cordeiro et al. [78] developed a deep learning model that was capable to anticipate the soil moisture availability in the agricultural land and addressing the sensor missing data and failure ambiguities. Among the studied algorithms, the k-nearest neighbors (KNN) algorithm bypasses the problems and accurately predicts the irrigation water need. Another study determined the soil moisture using gradient boosting with regression tree (GBRT) and found the best results for smart irrigation planning [79]. Abdulaziz et al. [80] utilized the Lagrange multiplier method in order to optimize the consumption of pumping power and water consumption. The model gives satisfactory results in optimizing the water consumption however, lacked in power utilization. The cited study [77] focused on the preprocessing of irrigation data and estimating plant growth using an adaptive neuro-fuzzy inference system (ANFIS) based on the soil, water level, temperature, and moisture conditions. The ANFIS proved to be optimum due to a precision value of 81% with an accuracy of 84.6% [77].
