Adaptive Control Design for a Three Phase Grid Connected Photovoltaic System with Experimental Analysis

Bhunia, Malay (2022) Adaptive Control Design for a Three Phase Grid Connected Photovoltaic System with Experimental Analysis. PhD thesis.

[img]PDF (Restricted upto 16/12/2024)
Restricted to Repository staff only



In view of meeting increasing power demand while minimizing carbon emission, more emphasis is being given to generate the power from renewable energy sources. Due to availability of abundant solar radiation, and low operational cost, the installation of PV based renewable power generation is increasing rapidly as compared to other renewable sources. The PV power generation can be operated either in standalone mode or grid connected mode. The control of grid connected PV power generation is challenging due to intermittent solar irradiance, and nonlinear environment dependent characteristics of PV source. The maximum power of a PV source depends upon the environmental parameters such as ambient temperature and irradiance. A Maximum Power Point Tracking (MPPT) algorithm is employed in PV control system to extract the maximum power from solar PV panels. The MPPT algorithm generates the reference PV voltage for PVVC. For single stage Grid Connected PV System (GCPVS), a PV Voltage Controller (PVVC) is employed to track the reference PV voltage, and simultaneously generates the reference grid current. Further, Grid Current Controller (GCC) tracks that reference grid current by controlling the switching of the inverter. To reduce the PV voltage ripple in single-stage GCPVS, DC-link capacitor is connected at the input of inverter. Also, coupling filters (such as Resistive Inductive (RL) filter) are used at the output of inverter to reduce the switching ripple in inverter output current. The performances of the PVVC and GCC depend upon the parameters of PV module, DC-link capacitor and RL filter. The internal resistance of PV module varies with its operating point resulting uncertainties in PV side. The DC-link capacitor varies due to aging. Coupling filter impedance varies with age, temperature and operating conditions. Further, grid current quality at low irradiance is an integral issue of GCPVS as all the passive components are designed at nominal power rating. To handle parametric uncertainties and intermittency in irradiance, it is necessary to design efficient adaptive controllers for GCPVS. The thesis focusses on design, development, and practical realization of different adaptive control schemes for achieving effective grid integration of single-stage three phase GCPVS. The thesis first develops an aging linked model of PV model by considering aging in PV parameters. Subsequently, a Quadrature Axis Small Signal Model (QASSM) of Grid Connected Inverter (GCI) with DC battery by considering the internal resistances, such as Equivalent Series Resistance (ESR) of DC link capacitor, internal resistance of PV cell, and internal resistance of coupling filter, is developed. This model is used to analyze the eigenvalues of the GCI. Subsequently, a QASSM of GCPVS is developed by considering GCI with PV source. From the eigenvalue analysis of the GCPVS , it is observed that internal resistance of the PV causes instability in GCPVS. Then the performances of GCPVS are evaluated at different aging periods. From this performance analysis, it is observed that PV voltage ripple, MPPT tracking efficiency and THD of grid current are deteriorated with increase in aging of GCPVS. Subsequently, the thesis develops different adaptive control schemes for GCPVS to handle the parametric uncertainties in GCPVS. First, a real-time parameter estimation based adaptive controller scheme, called as Self-tuning Sinusoidal Recursive Controller (STSRC), for GCPVS has been developed to handle the parametric uncertainties. The STSRC scheme employed an Improved Linear Sinusoidal Tracer (ILST) based Recursive Least Square (RLS) estimation called as IRLS. The IRLS estimates the GCPVS parameters effectively under distorted grid voltage profiles, and under low speed sampled data. The STSRC employed for both PVVC and GCCs. Efficacies of the proposed STSRC scheme are compared with Proportional Integral (PI) controller, Integral Sliding Mode Controller (ISMC) and Robust Nonlinear Adaptive Backsteeping Controller (RNBC). The comparison envisages that STSRC provides improved performance of GCPVS in terms of achieving low PV voltage ripple and low grid current THD. Then, a Lyapunov Based adaptive Voltage Controller (LBAC) is developed to control the PV voltage. The Total Harmonics Distortion (THD) of grid current increases at low irradiation. The THD of grid current not only depends upon the GCC but also depends upon the accurate tracking of PV voltage. Oscillation in PV voltage causes the distortion in grid current. This further degrades the grid power quality. Thus, tracking performance of PVVC plays an important role in better power extraction and in overall grid power quality. The LBAC is designed to provides critically damped PV voltage tracking with desired settling time despite the different disturbances in GCPVS. Efficacies of the proposed LBAC are compared with PI controller and RNBC. The comparison envisages that LBAC provides smooth critically damped PV voltage tracking with low voltage ripple as compared to PI controller and RNBC. The improved PV voltage tracking, yielded by LBAC, improves the THD of grid current in both high and low irradiance as compared to PI controller and RNBC. Finally, a cascaded Model Reference Adaptive Controller (MRAC) scheme is developed for GCPVS to handle the parametric uncertainties and disturbances. Apart from uncertainties, measurement noise affects the performances of the controller in GCPVS. However, use of additional filtering of feedback signals reduces the stability margins of controller. A cascaded MRACs are designed for both PVVC and GCC to achieve improved performances even under measurement noise. The proposed control scheme comprises of a reference model, and a Lyapunov based parameter adaptation scheme which provides the nominal tracking performance despite uncertainties in the GCPVS dynamics. Two separate reference models are chosen respectively for PVVC and GCC to define the critically damped desired tracking characteristics. The efficacies of the designed MRAC scheme are compared with PI controller, RNBC, LBAC and ISMC. Due to cascaded MRAC, the designed MRAC scheme is more noise resistant as compared to other controllers. Further, the MRAC scheme provides critically damped PV voltage response with reduced PV voltage ripple, and low grid current THD as compared to PI controller, ISMC, RNBC and LBAC. All the adaptive control schemes are first verified through simulation studies in MATLAB/SIMULINK environment. Then, efficacies of the different proposed adaptive control schemes are verified through real-time implementation using the 2.25 kW GCPVS prototype, developed in our laboratory. From the comparative analysis, it is observed that both LBAC and MRAC schemes exhibit critically damped PV voltage response with less ripple, and less grid current THD. However, MRAC scheme provides improved performance in robust PV voltage tracking with reduced grid current THD under the measurement noise as compared to PI controller, ISMC, RNBC and LBAC.

Item Type:Thesis (PhD)
Uncontrolled Keywords:Lyapunov Based Controller; Model Reference Adaptive Controller; Online Parameter Estimation; Single-stage Grid Connected Photovoltaic System; Small Signal Modeling.
Subjects:Engineering and Technology > Electrical Engineering > Power Systems > Renewable Energy
Engineering and Technology > Electrical Engineering > Power Systems
Engineering and Technology > Electrical Engineering > Power Transformers
Engineering and Technology > Electrical Engineering > Power Electronics
Divisions: Engineering and Technology > Department of Electrical Engineering
ID Code:10364
Deposited By:IR Staff BPCL
Deposited On:16 Dec 2022 02:04
Last Modified:16 Dec 2022 02:04
Supervisor(s):Subudhi, Bidyadhar and Ray, Pravat Kumar

Repository Staff Only: item control page