**1. Introduction**

As conventional energy sources are depleting day by day, the demand forrenewable energy sources is raising [1–3]. Solar energy sources are promising renewable energy sources for developed and developing nations due to being free, abundant, and environmentally friendly. Standalone photovoltaic (PV) systems for water-pumping applications are employed in remote areas [4,5]. Because of grid absence in remote places, standalone PV water pumping is installed for agricultural and household applications. Various electric motors have been used to drive the pumping system [6,7]. The DC motor-based pumping system requires maintenance because of commutator and brush presence. Therefore, DC motors are not frequently used for PV pumping applications. Single-phase induction motors have also been used for driving low-inertia torque load. Due to a complex control strategy, the induction motors are not efficient for pumping applications. Therefore, in this research work, a brushless DC (BLDC) motor has been considered as it has simple design control, low power range and requires maintenance-free operation compared to AC motors [8]. Distinct DC-DC converters

were contenders for optimizing PV module generated power with a soft-starting and controlling motor pump system [9–11]. The contemporary PV system has insubstantial converse competency. Therefore, Maximum Power Point Trackers (MPPT) is the indispensable constituents required for optimal power tracking from PV modules. Numerous MPPT methods have been occupied viz. Perturb and Observe (P&O), Increment Conductance (INC), Fraction Short/Open circuit etc. [12–14]. Under steady-state operating conditions, particular algorithms provide high outturn. However, these algorithms are found lacking under adverse weather conditions showing slow convergence velocity and being unable to achieve a global power point (GPP) for partial shading situations with high power oscillations around this point. Recently, different intelligent techniques viz. Fuzzy Logic Control (FLC) and Artificial Neural Network (ANN) have been employed for PV tracking [15]. However, because of complex fuzzy inference rules and individual sensor requirements, meta-heuristic algorithms have been employed recently. Genetic algorithms and artificial immune systems (AIS) are meta-heuristic algorithms used for non-linear stochastic problem solutions [16,17]. These algorithms are capable to resolve non-linear complication. However, due to a large population size and adaptive immune cell mechanism, the GA and AIS algorithms, respectively, have low velocity of convergence with large computational period. However, the implementation of selection, mutation and crossover process is complex with reduced convergence computational period. Currently, bio-inspired and swarm optimization have been derived as MPPT techniques. The particle swarm optimization is an evolutionary methodology based onthe nature of a swarm that can reduce oscillations around GPP [18]. The classical particle swarm optimization PSO technique has randomness in acceleration value with high regulation parameters as major problems. Nevertheless, variance of this algorithm is capitulated when randomness is miniaturized. Surrogating to swarm techniques, current bio-inspired algorithms viz. Firefly Algorithms (FA), Artificial Bee Colony (ABC), Cuckoo Search etc. are considered as bio-inspired MPPT and have the advantages of high convergence speed, and less transience with fast tracked performance [19–21]. Nevertheless, because of a lower number of bees, ABC technique has a slow velocity of convergence under fluctuating weather situations. Because of the deviating movement of a large number of fireflies, the response of the system becomes slow with the prerequisite high computational period. The Cuckoo search algorithm provides an efficient solution to non-linear problems. However, because of complex nest population and inadequate contingency, the Cuckoo technique has slow convergence velocity under varying environmental conditions. However, the implementation complexities with the tuning of parameters are amajor hindrance of this finding. The above-mentioned algorithms' drawbacks can be handled by applying Flower Pollination (FPA) as an MPPT technique. Ram et al. [22] has discussed the FPA algorithm for PV MPPT under dynamic operating conditions. This algorithm provides single-stage global searching, simpler coding, and lower tuned specification requirements with low cost implementation and has fast response compared to P&O and PSO techniques under dynamic weather conditions. Included in this work, a novel flower pollination algorithm is contemplated and associated with the hybrid ANFIS MPPT [23] algorithm. Compared to the FPA algorithm, the merits of the hybrid ANFIS-Flower Pollination Algorithm (FPA) are simple implementation, high convergence speed with tune parameters and easier code compilation. Due to presence of parasitic components, the voltage outcomes and power transform adequacy are restrained in classical switched-power converters. However, re-lift/triple lift methodology is employed by Luo converter to enhance the voltage balance and run-over the limiter issue. Equated with Zeta, single-ended primary-inductor converter (SEPIC) and Cuk converters, the Luo converter has accurate steady and dynamic system behavior. However, the Cuk converter comprises maximum transients and more settled periods with average system performance of SEPIC and Zeta compared to the Luo converter, which is applicable for upraised power utility and electrical drive operation [24,25]. In contrast with different employed power converters, modern Luo convertershave been considered for this research approach as they deliver better power/density ratio with economical implementation. The most recent attempt, using adaptive neuro-fuzzy inference system (ANFIS)-FPA-operated BLDC directed PV pump with advanced Luo

converter, has not been formerly conferred and examined using dSPACE (DS1104) platform under changing weather conditions.

#### **2. Complete System Formation**

Figure 1 illustrates the Luo converter-employed BLDC-driven PV pumping for a remote location. A hybrid ANFIS-FPA MPPT controller is operated to produce required pulse for power switched of Luo converter. This converter delivers better power/density ratio with economical implementation with interface between the inverter power circuit and solar system. Moreover, electronic commutation methodology controls voltage source inverter (VSI) employed BLDC motor in which winding current is adjusted with the help of a decoder in a proper sequence.

**Figure 1.** BLDC-driven Photovoltaic Complete System Formation.

#### *2.1. PV Generator*

In this research work, a two-diode PV cell model is considered (Figure 2) because it is a simple and accurate model compared to the single-diode PV cell. By means of photoelectric effect, the conversion of solar energy to electricity takes place and output power can be enhanced by connecting numerous solar cells in shunt or series as required. Both diodes are employed to represent polarization occurrence with current source exhibiting sun insolation, followed by power loss delivered by resistances (series/Parallel) used. The prognosis of the overall system is calculated based on accurate equivalent modeling. The output of the PV current is expressed mathematically as [26]:

$$I\_{PVo} = I\_{PVG} - I\_{RSC} \left(l' + 2\right) - \left(\frac{V\_{PVo} + I\_{PVo} \times R\_{series}}{R\_{Parallel}}\right) \tag{1}$$

where,

$$I' = \exp\left(\frac{V\_{PVo} + I\_{PVo} \times R\_{series}}{V\_{Thermal}}\right) + \exp\left(\frac{V\_{PVo} + I\_{PVo} \times R\_{series}}{A \times V\_{Thermal}}\right) \tag{2}$$

*IPVG* = Photo Current *IRSC* = Diode reverse saturation current *IPVo* = Output PV current *VPVo* = PV output voltage *Rseries* = Resistance in series

*RParallel* = Resistance in parallel *VThermal* = PV module thermal voltage *A* = Ideality constant of diode

**Figure 2.** Two diode PV cell model.
