Binary Hybrid Multilevel Inverter-Based Grid Integrated Solar Energy Conversion System With Damped SOGI Control

This paper presents a binary hybrid multilevel inverter (BHMLI) based grid-connected solar energy conversion system (SECS), controlled by damped second-order generalized integral (DSOGI). The BHMLI architect has a cascaded half-bridge array, which modifies the DC-link of the H-bridge of the voltage source inverter, and results in approximate reference waveform. It reduces the dV/dt of the H-bridge switches and improves output waveform quality. The DSOGI control damps the oscillations and overshoots and provides longer service period of low power switches at transient conditions. It is implemented in the multilevel inverter application for the first time in literature. The SECS is designed to inject active power to the grid, and also mitigates the harmonic and reactive power demands of the load. The cascading of ‘ $n$ ’ half-bridges and one H-bridge generate $(2^{(n+1)} - 1)$ output voltage levels. The maximum power extraction from solar photovoltaic (PV) array is achieved through incremental conductance (IC) algorithm based maximum point tracking (MPPT) operation of the DC-DC converter. An isolated single-input multiple-output single-ended primary inductance converter (SIMO-SEPIC) realizes it. SECS with 15-level BHMLI is analyzed with extensive simulation as well as with a hardware prototype. Moreover, the shunt active filter functionality of the system at various load conditions are verified and maintained the grid power quality within the IEEE-519 standard throughout the operation. The laboratory developed setup is tested for 5 kW, 400 V, three-phase system, and the experimental analysis at steady-state and dynamic variations of load-side and insolation variations validate the theoretical claims.


I. INTRODUCTION
The rapid growth of electrical energy demand as well as the requirement of clean and free energy for environmental safety, give a spotlight to the renewable energy (RE) based power generation, and it is increasing globally at a fast scale. The solar photovoltaic (PV) systems are gaining more attraction among grid integrated RE technologies [1]. The advantages such as long life, ease of installation, rooftop applications, free and abundant energy, clean energy production, noiseless operation, absence of moving parts and The associate editor coordinating the review of this manuscript and approving it for publication was Zhilei Yao . less maintenance are the vital advantages of solar PV. The efficient power transfer from the solar PV to the distribution grid with is the modern research trend.
The grid-connected solar inverters avoid bulky battery banks and save space and cost. A conventional two-level (2L) inverter-based solar energy conversion system (SECS) is reported in [2], where a controller of modified proportional-resonant (M-PR) is operating both in stand-alone as well as grid-connected modes. The primary advantages are circuit and switching simplicity. The peak-peak current ripple in the 2L inverter is analyzed and compared with the multilevel inverter (MLI) in [3], and some pulse width modulation (PWM) schemes for reducing the current ripple is also reported. The higher switching stress limits the application of 2L to low power. Moreover, the higher switching frequency and complex filters reduce the efficiency and introduce electromagnetic interference (EMI) issues to nearer power and communication lines.
The grid-synchronization part of SECS is carried out by voltage source inverters (VSI), and the MLIs are introduced to overcome the demerits of 2L inverters. These are classified as neutral point clamped (NPC), flying capacitor clamped (FCC) and cascaded H-bridge (CHB). The features and power loss distribution of a three-level (3L) stacked neutral point clamped MLI is analyzed for the first time with solar PV applications in [4]. However, the THD is reduced, but the topology needs a larger number of switches for a 3L output. The literature in [5] addresses the unbalanced fault conditions of a 3L-NPC, during a fault, it generates errors and introduces a delay in the power system because of the intricate design or the neglected negative-sequence component of the grid. An FCC-MLI is reported in [6], [7] where the study focuses on the control side of the grid-connected power conditioning unit. The number of capacitors and semiconductor devices increase the cost, size and complexity. The capacitor voltage unbalancing is another issue. Hence it is less appealing to find the application for FCC-MLI. The CHB inverters are a prevalent MLI, and it provides a highly modular topological structure and higher power transfer capabilities. However, it requires a large number of switches for higher output voltage levels and isolated DC sources. It is well used for PV applications and reported in [8], and it is in conjunction with a SEPIC converter. The power and energy balance among the H-bridges of a CHB-MLI are crucial, and various PWM techniques are reported recently by addressing that issue [9]. Numerous as good as ever MLI topologies are now and again reporting these days for PV applications. The principle fascination is that the solar panel management acknowledges the demand for isolated DC sources in MLIs. Kamal Al Haddad et al. reports the packed-U-cell MLI topology, and it offers high-energy conversion quality using a lesser number of capacitors and power switches. It is modified and reported for PV applications in [10]. The paper [11] is an MLI based on two isolated 2L inverters, which are connected to the grid through a three-phase transformer having open-end windings on the inverter side, realizes a 3L-MLI. A nine-level, grid-connected inverter for PV application, based on two CHBs with different dc-link voltages is reported in [12]. An additional FCC-MLI structure is necessary for this topology, and this will prompt unequal voltage levels. Therefore, a supplementary voltage balancing circuit is essential, that results in over component count. A modified cascaded grid-connected MLI and the topology comprises of a bidirectional switch included to the standard CHB-MLI [13]. It adds additional switching and conduction losses and also increases the circuit complexity. An asymmetrical MLI presented for operating under wide dc-link voltage variation consists of a level doubling network connected in series with a full-bridge [14]. The interconnection of such units increases the number of output voltage levels. An improved CHB-MLI is reported in [15], by reducing the leakage currents and hipe the efficiency, and it shows low switching and conduction losses. An asymmetric CHB-MLI based on trinary DC sources is reported in [16]. It is a single-sourced 27-level MLI. A modular multilevel single-delta bridge-cell inverter is reported in [17] intended for utility-scale PV systems. An asymmetrical cascaded MLI with individual DC-DC converters fed from asymmetric power PV panels is published in [18]. The separate maximum power point tracking (MPPT) algorithms are implemented for different panels. However, the output voltage levels are high, but the complex control will slow down the dynamic performance of the system. A three-level step-up converter for improving the power conversion efficiency by lowering the voltage stress as well as balancing of the dc-link capacitor voltages is connected to the three-level inverter is presented in [3]. Additional voltage balancing is implemented, and the MLI is incapable of self-regulating the DC bus. The modified H-bridge-based MLI given in [19], includes two legs, a usual two-level leg and a T-type leg. The T-type is a modified 2L leg but connected to the split DC-link mid-point through a bidirectional switch. An asymmetrical CHB-MLI reported in [20] has the DC source asymmetry generated through a high-frequency DC-DC converter, and this will increase the overall component count, conduction losses and system complexity. A hybrid nine-level MLI [21] uses a PWM technique for nine-level voltage output. The reduced switch count is the main attraction of the topology, ten switches for nine levels. The globally accepted two stage SECS has more stability than single stage SECSs, because, the maximum power extraction and DC-link voltage stabilization are done by DC-DC converter part. Therefore, it is necessary to use an efficient, flexible and simple DC-DC converter to interface between PV and VSI.
A buck-boost DC-DC converter is presented in ref. [22] for PV applications. The wide range of voltage regulation capability is the advantage. It needs only one buck-boost stage for voltage regulation by employing the reconfiguration of the switching stage with a smooth transition between the modes. A new class of converters based on capacitor link DC-DC converters-Cuk, SEPIC, and Zeta converters-that operates under the critical conduction mode has been reported [23]. An auxiliary circuit is used to get zero-current and zero-voltage switching for all the switches during the transitions. The value of dv/dt and EMI issues and voltage ringing across the switches are reduced. A high gain DC-DC converter is reported by combining the switched capacitor and regenerative boost configuration [24], but the device count is increased. A control scheme for the boost converter operated with a grid-connected PV system is reported in [25]. The MPPT, as well as limited power point tracking control schemes, are employed in the boost converter. Since the inability of the boost converter to withstand wide range duty ratio is not addressed. A high gain DC-DC converter with the coupled inductor and intermediate capacitors VOLUME 8, 2020 are used to achieve high gain [26]. To recover the leakage energy and to alleviate the voltage spikes, a clamped circuit is used. A review is reported with differential power processing converter technology in [27], and the ability of SEPIC converter for wide duty ratio control and acceptance in SECS technology is also discussed. The MPPT operation of DC-DC converter requires flexibility to buck and boost operation and must be stable at wide duty ratios.
The traditional hill climb algorithm based MPPTs such as perturb and observe (P&O) and incremental conductance (IC) algorithms provide better results at steady state. Still, the dynamic response at partial shading conditions is inadequate [28]. It has been improved by adaptive techniques such as fuzzy logic control, particle swarm optimization, artificial neural network, etc. give accurate MPPT during adverse insolation conditions. The genetic algorithm, as well as some evolved techniques, exhibits superior performance and also have the facility to add or remove constraints [29]. But these complex algorithms reduce the speed of a larger system. Therefore, for a large grid-connected solar energy conversion systems are using modified accurate hill-climbing techniques to improve the overall dynamic performances [30].
Another inevitable part of the SECS is grid synchronization control. A linear quadratic regulator based solar energy conversion system is reported in [31]. The adaptive sliding mode control scheme is reported in the literature for current control in grid integration of PV system [32]. It is less accurate, and power calculation steps are tedious. A second-order general integral (SOGI) based phase-locked-loop is presented in [33], and it shows poor damping at oscillations and overshoots. Therefore, it causes component failure. A two-stage solar PV system with a damped second-order generalized integral (DSOGI) algorithm is reported in [34]. The fundamental active power components of load currents are extracted for evaluating the reference grid currents. By using the appropriate damping factor, the grid current-control eliminates the overshoots and provides a steady response. Moreover, it gives a fast and accurate operation.
In this paper, the DSOGI controller is implemented to control a binary hybrid MLI by which real power control, reactive power control, harmonic elimination and source current balancing are verified. This paper is organized in such a way that the solar energy conversion system is described in Section II. The control part is explained in Section VI followed by the modeling of SECS in Section III. The simulation results and experimental validation are given in Section VII and Section VIII, then the conclusions are in Section IX.

II. CONFIGURATION OF GRID CONNECTED SECS
The configuration of the three-phase solar energy conversion system is shown in Fig. 1. A nonlinear, unbalanced load is connected at the point of common coupling (PCC). An RC ripple filter to eliminate the voltage ripples due to MLI switching is also connected at the PCC. This paper reports a fifteen-level binary hybrid multilevel inverter in SECS. It has an array of three half-bridge converters (hBC), feeding from  isolated DC-sources V 1 , V 2 and V 3 , in the ratio of 1 : 2 : 4 respectively. The operation of hBC switches S a1 , S a1 , S a2 , S a2 , S a3 and S a3 is modulated for unipolar stepped voltage at the H-bridge input of phase-a is V Ha (or with low switching frequency level shifted pulse width modulation), as shown in Fig. 2 (a). The H-bridge switches H a1 − H a4 switched at the fundamental frequency, and produce desired AC output voltage, which is depicted in Fig. 2 (b).
The centralised solar PV field is feeding SIMO-SEPIC for three-phase balanced power injection, it is designed to provide isolated DC sources for the hBCs in each phase, and it controlled by IC algorithm based MPPT. T a , L a and C a are semiconductor switch, primary side inductance and primary side capacitors of the SIMO-SEPIC, respectively. The multi-winding transformer is designed with a turns ratio of 4 : 1 : 2 : 4.
The relation between the output voltage levels and number of half-bridges is derived as, where, n is the number of hBCs in phase-a. The number of power electronic switches in one phase is determined as, The positive voltage levels in the output voltage of phase-a V ao are, V 1 , 2V 1 , 3V 1 , 4V 1 , 5V 1 , 6V 1 and 7V 1 . Similar negative voltage levels and a zero level are contained in a complete cycle of output voltage as depicted in Fig. 2 (b). The fundamental output voltage waveform is shown in the diagram, and the expression is derived by Fourier series analysis. Due to the quarter wave symmetry along the x-axis, both Fourier coefficients A 0 and A n are zero [38]. The expression for fundamental output voltage is, where, S f is the switching factor, calculated by using MATLAB Simulink Mathfunction as, V 1 is the output voltage step magnitude and α 1 to α 7 are the switching angles, as depicted in Fig. 2.
The system efficiency is calculated by considering the power losses P L as, P out = P in − P L P L (total) = P sl + P cl + P osl + P gl (5) where, P Ltotal is the total power loss, P sl is the switching losses, P cl is the conduction losses; P osl is the OFF state losses and P gl is the gate losses. The upper hBC switches, S a1 , S a2 and S a3 are operated to introduce the respective DC sources to the output, and the lower hBC switches S a1 , S a2 and S a3 are operated to circumvent the respective DC sources from the output voltage waveform. The upper and lower switches of single hBC are complimentary in operation. The output voltage levels are considered as modes for the explanation. Some of such operating modes, Mode-I of 0V 1 output, Mode-VI of 5V 1 and Mode-VII of 7V 1 are delineated in Figs. 3 (a)-(c). Moreover, the H-bridge switches are controlled by zero-crossing detection of the fundamental frequency reference waveform. The overall switching combinations for the fifteen discrete output voltage levels are depicted in TABLE 1.

III. SECS DESIGN
Voltage rating of n th hBC IGBT switch is, where, V n is the DC-link voltage of n th hBC and V m is the maximum overshoot allowed (10 % of V n ). Three half bridges, with isolated DC-sources in the ratio of 1 : 2 : 4 are cascaded as shown in Fig. 3. Therefore, the voltage rating of H-bridge switch V H is, where, V mH is, 10 % of 7V 1 . Current rating of n th hBC IGBT switch is, where, I SECSmp is the peak of maximum output current through the one phase of the SECS. I r is the maximum ripple current allowed (5 %). Where, I SECSmp is the sum of active and reactive currents (maximum possible with this practical inverter). VOLUME 8, 2020 The multiplication factor 1.25 is used in (8) to incorporate the safety factor. Current rating of H-bridge switches are same as that of hBC switches.
The DC-link capacitors C 1 , C 2 and C 3 at input of hBCs (generally call C n ) are designed as [39], where, V no is the rms value of n th hBC output voltage, a is the overloading factor (1.2), t time to be which the DC-link voltage to be recovered, V nmax is maximum possible and V nmin is the minimum DC-link voltage of n th hBC. The magnitude of interfacing inductance L i is designed based on the active power rating of SECS as, The magnitude of δ is selected by the active power sharing, and it is evaluated by the control algorithm. P SECS is the active power rating of the SECS and ω is the fundamental grid frequency in rad/s. Voltage relation of the SIMO-SEPIC is derived as, , D is the switching duty ratio of switch T a and V PV is the input voltage across C PV .
The inductance value of L a is derived as [8], where, D max is considered as 0.8 at continuous conduction mode, and the isolating transformer inductance also reduces the current ripple, therefore, to get a minimum possible L a , I PV is considered as 15 % of I PV . f ss is the switching frequency of T a . V PVmin is the minimum PV voltage at which converter operates satisfactorily to generate isolated outputs. where, and N s is the total number of secondary windings and N p is the total number of primary windings. V sec min is the minimum value of cumulative DC-link voltages, 7V 1min . V 1min is derived from the relation between the fundamental grid voltage (v sab ) and inverter output voltage derived in (3).
To acheve, power transfer from SECS to the grid PCC, The factor 1.1 is considered to account voltage drops at various parts of the SECS as well as to maintain, V ao to be higher than V sa . The output voltage across the n th secondary winding of SIMO-SEPIC V n is derived as [8], (18) where, N p is the primary and N sn is the n th secondary turns of multi winding transformer. Individual DC-link voltages generated by SIMO-SEPIC at the out terminals are derived from (18) as, The design of T a is, where, I o is the output current of the converter, The Diode in the secondary side of the winding also has the same voltage and current ratings of IGBT switch [11].

IV. EFFICIENCY CALCULATION
The efficiency of the BHMLI is evaluated by using (25) η BHMLI = P in − P loss P in (25) where, P in is the total input power of BHMLI. P loss is the total loss of every semiconductor devices during conduction (P lc ), blocking (P loff ) and switching (P ls ) states. Where, P loff is due to reverse leakage current, and it is practically negligible, and therefore, loss associated with leakage current is also omitted [35]. The instantaneous conduction losses through the diode and transistor are evaluated as [36],  device at an instant is same as the instantaneous magnitude of output current (i SECS (t)). The average conduction loss is estimated from (26) and (27) as, where, h S (t) and h D (t) are the instantaneous number of IGBT switches and number of diodes in conduction. The total switching loss is calculated by summing up the individual switching loss of each switch from  bypassed. According to equation (6), the hBC switch S a1 and S a1 are rated at 1.1V 1 , S a2 and S a2 are rated at 2.2V 1 , S a3 and S a3 are rated at 4.4V 1 and H-bridge switches at 7.7V 1 . Therefore, if a voltage transient appeared at t with magnitude +6V 1 as depicted in Fig. 4 (b), hBC switches damages, and leads to overall system failure. Similar analysis can be carried out in current states also. Hence, a sophisticated, perfectly damped control algorithm is required to eliminate the overshoots and oscillations in voltages and currents at load side, grid side and PV side fluctuations. A DSOGI control reported for conventional two level inverter in [34], is implemented on the BHMLI for satisfactory SECS operation.

A. BHMLI CONTROL USING DAMPED SOGI
The PV power available at the BHMLI input is transferred to the utility side by the damped second-order generalized integral based control [34]. It is designed to deliver active power to the grid from solar PV panels and also mitigate harmonic as well as reactive power requirements of load and thereby keeping the grid power quality standards. The overall control block diagram is depicted in Fig. 5. The two-phase quantities are sensed from the three-phase three-wire system, and the third phase parameters are estimated to reduce the number of sensors. Two line-line grid voltages v sab and v sbc , load currents i La and i Lb and the grid currents i sa and i sb are sensed.
The three-phase grid phase-voltages are estimated from line-voltages which are passed through individual bandpass filters tuned at 314 rad/s as, The load current and grid current os phase−c are estimated as, The amplitude of average phase voltage is estimated as, VOLUME 8, 2020 The unit in-phase templates are generated from (30) and (32) as, The unit quadrature components are also estimated as, The intermediate DSOGI control signals are evaluated as [34], The constants, k D , k l and ρ are calculated by Routh Hurwitz (RH) stability analysis of the control transfer function, and it is derived as, where the Zero-order hold is transformed into 1−e −Ts s and the unit template of the sine wave is transformed to 1 s 2 +1 . Where ω =314. Damping factor ρ is selected as 0.5. According to RH criterion of stability, The characteristic equation (CE) satisfy necessary conditions such as no missing terms and no sign changes. The stability analysis of the DSOGI controller is evaluated and compared with the traditional SOGI algorithm. It is found that the poles and zeros are aligned at the left-hand side of the imaginary axis in the pole-zero plot. As it is a frequency fixed tuning algorithm, the slight allowable variations in the grid frequencies are accommodated by spreading the bandwidth accordingly. Therefore, the selection of damping factor must be a trade-off between tuning capacity, speed of response, overshoot-damping and flexibility with frequency variation.
The performance of the DSOGI control algorithm for the fundamental load component extraction of load current at ω = 314 rad/s is evaluated by bode plot analysis, and it provides higher attenuation below and above ω. Hence, fine fundamental component tuning is ensured.
The fundamental load current component i Lfa is passed through the sample and hold circuit, and it is triggered by the zero-crossing pulses of quadrature unit templates of grid voltage. The absolute value gives the active component of the load current. The total active load current component is evaluated as, The reference of individual grid current amplitudes are estimated as, where, I * PV is the DC-link regulation output, derived as [32], k pr , k ir are the proportional and integral constants of PI controller. V * 1 is the reference DC-link voltage evaluated 37220 VOLUME 8, 2020 from (17), and I pf is the PV feed-forward used to improve the dynamic performance of the system, and it enhances faster PI control operation. It is derived as, P PV is the output PV power, The reference grid currents are estimated as, i sa = I * s × u pa ; i sb = I * s × u pb ; i sc = I * s × u pc (48) It is compared with the sensed grid currents, and the error is passed through a current controller to produce three-phase inverter switching references i * sae , i * sbe and i * sce respectively. It is fed to gate pulse generation unit, and the BHMLI gate pulses are produced.
In order to calculate the switching angles of the MLI, α k is the k th switching angle as shown in Fig. 2.
The selective harmonic elimination and other complicated switching angle calculation algorithms are not used here. Because, sometimes, the SECS is required to inject harmonics and reactive power to the PCC for the considered load. Seven carrier signals are used, and compared with the reference signal. Current references 1 to 7 are the constants with amplitudes 0.5 V 1 , 1.5 V 1 , 2.5 V 1 , 3.5 V 1 , 4.5 V 1 , 5.5 V 1 and 6.5 V 1 respectively. The crossing point of reference signal and the carrier signal in the x-axis is evaluated using MATLAB Simulink relational operator. The corresponding switching pulses with angles α 1 to α 7 are feed into the respective switches (TABLE 1). This quarter wave switching pulse generation program is continued in the entire cycle, and it is the switching angle calculation method.

B. MPPT CONTROL
The efficient PV power transfer is achieved in the SECS by using incremental conductance algorithm based MPPT controller. It is comparatively simple and fast [34]. Power drawn from the PV panel, P pv is calculated by sensing PV voltage v pv and PV current i pv as given in (47). At the maximum power point (MPP), The conductance, The IC of PV is g ac = |di pv | |dv pv | . But, the PV absolute conductance, g DC = i pv v pv [34]. From the equation (50), The IC algorithm evaluates the MPP operating pint and produces the duty cycle D as output. The PWM with a triangular carrier wave gives switching pulses to the SEPIC switches T a , T b , and T c .

VII. SIMULATION RESULTS AND DISCUSSIONS
The simulation analysis of this system is carried out in MATLAB/Simulink platform. The system parameters are depicted in Appendix A. The BHMLI generates 15-level output voltage waveforms, as shown in Fig. 6. The three-phase voltage output waveform of SECS has the peak voltage as 325 V, and the frequency is 50 Hz. The system is designed in a way that, the SECS supporting the nonlinear load demand, connected at the PCC and the remaining power is delivered to grid at rated insolation (1000 W /m 2 ). Fig. 7 shows the waveforms of irradiation level I rr , PV output voltage V PV , PV output current I PV , output PV power P PV , three-phase grid voltages v s , grid currents i s , load currents i L and SECS currents i SECS at various operating conditions in single frame, and the detailing is as follows.

A. STEADY STATE OPERATION OF SECS WITH BALANCED NONLINEAR LOAD AT CONSTANT IRRADIATION
The steady-state behaviour of the SECS is analysed with applying a balanced non-linear load at the PCC. The irradiation level is kept at 1000 W /m 2 , and the waveforms are depicted before 1.98 s of Fig. 7. SECS feeding power to load and the grid. Therefore, the grid currents are 180 0 out-of-phase from the phase-voltages. One can witness that the grid currents are maintained sinusoidal and balanced. Therefore, the steady-state operation of SECS is found to be satisfactory.

B. DYNAMIC PERFORMANCE OF SECS WITH UNBALANCED LOAD CONDITION
A load unbalance introduced by thrown-off of the lad at phase−a at 1.98 s as shown in Fig. 7. During this, load transition, P PV remains constant. Therefore, the balanced VOLUME 8, 2020 grid currents are increased. The load single-phasing is made ineffective in the grid by unbalanced operation of SECS; therefore, i SECS are unbalanced. The grid current increases suddenly and attains the steady-state within a cycle duration. The sudden load removal does not create any overshoots and oscillations in any part of the system, and it validates the effect of the damped controller. The fast, steady-state achievement of the system parameters are also witnessed in this analysis. Similar satisfactory performance is also observed during load restoration at 2.15 s. Moreover, the grid currents and voltages are balanced and sinusoidal.

C. PERFORMANCE OF SECS UNDER VARIABLE IRRADIATIONS
The irradiation level is changed to observe the system behaviour. For that, at 2.05 s, the solar irradiation is reduced from 1000 to 500 W /m 2 . It affects V PV , but it maintained nearly constant. I PV reduced to nearly half, and therefore, it is reflected in P PV . The insolation variation with unbalanced load reduces the SECS output, and it is witnessed in the waveforms of i SECS . As the grid is accepting power from SECS, it also reduced and settled nearly half of the rated. One can witness it in i s . Further, the irradiation is reduced to zero-level to examine the system performance at unbalanced load at 2.1 s, as shown in Fig. 7. AS I rr reaches zero, I PV is reduced to zero, but there is a small value V PV can be observed due to charged capacitor C PV . P PV becomes zero, instantly, and it reflected in i SESC . It fails to mitigate the active demand of unbalanced nonlinear load, and it is drawn from the grid. During these transitions, the system performance is satisfactory, and the speed of steady-state attainment is also found satisfactory. Moreover, during the unbalanced, zero-insolation condition, the SECS performs as an active shunt filter and provides balanced sinusoidal and in-phase grid currents with the respective phasevoltages.
The harmonic analysis of the system parameters is depicted in Fig. 8. Fig. 8 (a) shows the harmonic spectrum of load current, and the THD is 24.3 % with 7.52 A fundamental amplitude. The current injected to the grid from the SECS has a THD of 1.3 % with amplitude 50.8 A fundamental as shown in Fig. 8 (b). The grid current and voltage harmonic magnitudes are maintained within the IEEE-519 standard. Therefore, the theoretical claim of the BHMLI based grid-connected SECS for power quality improvement is validated.
As a summary, the simulation analysis of the system provides satisfactory steady-state and dynamic performances at various load and irradiation conditions.

VIII. EXPERIMENTAL VALIDATION
A hardware prototype of the solar energy conversion system is developed in the laboratory with 5 kW, 400 V, 50 Hz three-phase grid-connected solar energy conversion system. SECS provides active power to the grid after meeting the load demands. The 15-level BHMLI is developed by using IGBT switches with protection driver., and the DSOGI and MPPT controls are realized by using dSPACE Micro Lab Box-1202. The waveforms are observed in digital storage oscilloscope (DSO), scope meter and power quality (PQ) analyzer. The design values of the system parameters are depicted in Appendix B.
The BHMLI output voltage waveforms of the experimental set are depicted in Fig. 9. The uni-polar output voltage waveform at the H-bridge input V Ha is shown in Fig. 9 (a) and the fifteen level output voltage waveform is depicted in Fig. 9 (b). The performance of SECS is analyzed for various load side and irradiation variations, and the results are depicted in the following sub-sections.

A. INTERMEDIATE SIGNALS OF DSOGI BASED GRID CONTROL
The intermediate signals of the DSOGI based control algorithm with a nonlinear load is shown in Fig. 10. The load current waveform i La , the derived signals i La1 , i La2 and i La3 are given in Fig. 10 (a). Fig. 10 (b) depicts the waveforms of i La4 , i La5 , i La6 and i Lfa . The grid voltage V sa , load current i La , grid current reference i * sa and the actual grid current i sa are shown in Fig. 10 (c). From the figures, one can witness the sinusoidal fundamental grid current reference i * sa , with same phase that of v sa .

B. STEADY STATE PERFORMANCE OF SECS WITH LINEAR BALANCED LOAD
At constant irradiation and a balanced linear local load, the performance of SECS is analyzed, and the results are depicted in Fig. 11. The SECS delivers 4.82 kW active power, and the load draws 2.7 kW. The remaining power is fed to the grid without altering the grid power qualities. The voltage and current waveforms of the grid during the SECS operation is expressed in Fig. 11 (a). Fig. 11 (b) shows the linear load current of phase−a and the SECS current waveform is in Fig. 11 (c). The three-phase power injected to the grid by the SECS is shown in Fig. 11 (d). Load power is displayed in Fig. 11 (e), and Fig. 11 (f) demonstrate the SECS power. From these test results, it is clear that the performance of SECS with a linear balanced load is found satisfactory. Moreover, it keeps the grid power quality standards.  The three-phase balanced, active power is injected to the grid at unity power factor, and zero reactive power is explored in Fig. 12 (j). The single-phase load power is given in Fig. 12 (k) and the grid current harmonics limit is within the IEEE-519 standard [40], as witnessed in Fig. 12 (l). Therefore, the performance of the system at the unbalanced load condition is verified.

D. PERFORMANCE OF SECS UNDER BALANCED NON-LINEAR LOAD
The test results of SECS at balanced non-linear load condition are depicted in Fig. 13. Figs. 13 (a)-(c) show the waveforms of v sab with i sa , i La and i SECS . It is realized that the grid current is sinusoidal and in phase with the grid phase voltage. The three-phase grid power, load power and SECS power are shown in Figs. 13 (d)-(f). It shows the unity power factor operation in grid power. SECS delivers power to the grid and the non-linear load. Figs. 13 (g)-(i) depict the harmonic spectra of i sa , i La and v sab . From the results, it is clear that the SECS mitigates the non-linear load with 24.7 % THD and the grid current and voltage THDs are maintained within the IEEE-519 standards [40].  Figs. 14 (a)-(c) that the three-phase grid currents are balanced, sinusoidal and steady at non-linear unbalanced load. The load current waveforms are depicted in Figs. 14 (d)-(f). The three-phase load unbalancing is realized by opening one of the load phases. The unbalanced operation of the SECS for grid current balancing is visualized in Figs. 14 (g)-(i). The three-phase balanced, active power injected to the grid at a unity power factor, as shown in Fig. 14 (j). The single-phase load power is given in Fig. 14 (k) and the grid THD is within the five percentage limit [40], as shown in Fig. 14 (l).

F. PERFORMANCE OF SIMO-SEPIC
The Solar PV field is directly coupled to the SEPIC converter, which is controlled for tracking MPP. The three isolated DC outputs voltages and currents of the SEPIC V 1 , I 1 , V 2 , I 2 , V 3 and I 3 are shown in Figs. 15 (a)-(c). Individual output power through the output ports of SIMO-SEPIC P 1 , P 2 and P 3 are shown in Figs. 15 (d)-(f). The PV output voltage and current waveforms V PV and I PV are depicted in Fig. 15 (g), and power is shown in Fig. 15 (h).

G. DYNAMIC PERFORMANCE OF SECS UNDER SUDDEN IRRADIATION VARIATIONS
The performance of SECS is tested at sudden insolation variations, and the results with waveforms of V PV , I PV , V 1 , V 2 , v sa , i sa , i sb and i sc are shown in Fig. 16. The sudden irradiation fall with constant load condition reduces the PV power, and the corresponding dip is visible in I pv . But the SIMO-SEPIC output voltages V 1 and V 2 are maintained nearly constant, which is presented in Fig. 16 (a). The reduction in three-phase balanced grid current injection is witnessed along with the grid voltage v sa in Fig. 16 (b). The test of SECS with sudden irradiation rise is carried out, and the corresponding variations are shown in Fig. 16 (c). As the PV power increases with irradiation rise, the increment in the three-phase balanced grid current is visible in Fig. 16 (d). Therefore, during sudden irradiation variations, the DC-link voltage is maintained at steady, and the overshoots and oscillations in grid parameters are eliminated, it also counted as the merit of SIMO-SEPIC controlled by IC-MPPT.

H. DYNAMIC PERFORMANCE OF SECS AT SUDDEN LOAD VARIATIONS
The dynamic behaviour of SECS with a sudden load variation is tested and the waveforms of v sa , i La , i Lb and i Lc , i sa , i sb and i sc are depicted in Fig. 17. One phase of the three-phase nonlinear is disconnected suddenly, and the load current waveforms with a grid phase-voltage are shown in Fig. 17 (a). With sudden load single phasing, load-power is dropped. Then that much power is injected to the grid. One can witness a balanced and in-phase grid currents and its variations in Fig. 17 (b). The waveforms of sudden load restoration with a grid phase-voltage are shown in Fig. 17 (c) and the corresponding grid power decrement is visible in Fig. 17 (d).
During the transient conditions, the SECS operation is found to be satisfactory, and the peak overshoots are eliminated without compromising the speed of operation by the control algorithm. It is the evidence of damped SOGI performance in BHMLI for component safety.
The photograph with descriptions of the experimental setup developed in the laboratory is shown in Fig. 18. A dSPACE Micro-Lab-Box is used for realizing the control part, and it is marked as 1 in Fig. 18. 2 is a solar array simulator with 10 kW capacity, it has seven individual panels, 3 is digital storage oscilloscope, used to evaluate the gate pulses and internal control signals. The dSPACE is controlled through a MATLAB interfacing PC, which is marked as 4. 5 is a multimeter, 6 is the power supply for the switch protection driver and 7, 14 and 15 are the IGBT modules, which is the   part of BHMLI. 8 is the three-phase autotransformer, used as distribution grid. 9 and 10 are power quality analyzer and power logger, used to evaluate the grid power quality and parameter readings. 11 is a scope meter with isolated channels used to catch the dynamic performance of the system. 12 is the SIMO-SEPIC DC-DC converter with isolating transformer and 13 is the optocoupler driver circuits for the gate pulses. 16 is an uncontrolled diode bridge rectifier used as a nonlinear load in the experiment, and 17 and 18 are the variable resistive and inductive loads connected at the DC-side of the rectifier. 19 and 20 are the variable inductive and resistive load used as a linear load. 21 is the interfacing inductor, 22 is the RC ripple filter connected at the PCC, 23 and 24 are the voltage and current sensors used for the control, and its supply unit is in 25.

IX. CONCLUSION
A grid-connected solar energy conversion system, based on a binary hybrid multilevel inverter is designed and tested in this literature. The fifteen-level BHMLI is realized by ten semiconductor switches and three binary weighted isolated DC-sources. The advantages of the reduced component count, reduced device rating and lower switching frequency are validated in this work. The acceptability of damped SOGI control in BHMLI for eliminating the overshoots and oscillations during load side fluctuations are tested and found satisfactory. Therefore, the low rated components are made safer in the adverse operating conditions. The single input multiple output SEPICs in each phase, connected to the common PV field realizes a balanced power transfer to the grid. It also provided the isolated DC-sources for BHMLI operation and stabilized the DC-link voltages during sudden PV output variations. Moreover, it achieves a fast, maximum power point tracking during the dynamic insolation variations by using incremental conductance algorithm. The extensive analysis in the experimental setup proves the satisfactory SECS integration to the grid during adverse operating conditions. It mitigates the harmonic demand and active power of load and validates the shunt achieve filter functionality during the entire operation. A balanced, sinusoidal and in-phase grid currents are achieved, at dynamic load and PV variations. Therefore, SECS operation maintains the grid power quality within the IEEE-519 standard. He has worked as Research Associate with the Integrated Rural Technology Centre, Palakkad, and as an Engineer with National Hydro Electric Power Corporation, Faridabad, India, from 1997 to 1998. He joined the Department, in 1999, and became an Associate Professor, in 2009. He has worked with Technical University, Dresden, Germany, for one year under DAAD scheme, in 2003. He is currently a Professor with the Department of Electrical and Electronics Engineering, Government College of Engineering, Kannur, Kerala, India. His research interests include power quality, power electronics applications to power systems, and renewable energy.
Dr. Jayaprakash is a Life Member of the Indian Society for Technical Education (ISTE). He has a teaching experience of 15 years. He is serving as a Reviewer for many IEEE and IET Transactions and conferences. He is also serving as an Advisor for the IEEE IAS Chapter, Government College of Engineering, Kannur. He has 18 journal publications and about 60 conference proceedings of reputed national/international conferences.
UMASHANKAR SUBRAMANIAM (Senior Member, IEEE) worked as an Associate Professor, the Head of the VIT Vellore, a Senior R&D, and a Senior Application Engineer in the field of power electronics, renewable energy, and electrical drives. He is currently with the Renewable Energy Lab, College of Engineering, Prince Sultan University, Saudi Arabia. He has more than 15 years of teaching, research, and industrial R&D experience. He has published more than 250 research articles in national and international journals and conferences. He has also authored/coauthored/contributed 12 books/chapters and 12 technical articles on power electronics applications in renewable energy and allied areas.
He  He is also serving as the Director of the Science and Technology Unit and the Leader of the Renewable Energy Laboratory, PSU. He has authored many published articles in the area of control systems. His research interests include the hardware implementation of control theory, signal processing, networked control systems, nonlinear control design, unmanned aerial vehicle (UAV), and renewable energy. He has served as a Reviewer for many journals, including the IEEE TRANSACTIONS ON