Bag, Aurobinda (2018) Development of Adaptive Controllers for Grid Integration of a Three-Phase PV System. PhD thesis.
|PDF (Restricted upto 25/02/2021) |
Restricted to Repository staff only
A PV system, particularly a large power one is usually operated in a grid tied mode.In view of achieving effective grid integration of a PV system, the necessity of important control algorithms are as follows. At first, Maximum power point tracking control for extraction of maximum power from the PV panels. Then in order to control the transfer of active power and reactive power from the inverter and grid to load, it is necessary to estimate the reference inverter current. Thus, there is a need of employing an adaptive filtering algorithm for this estimation. Furthermore, an appropriate switching signals for the inverter to be generated to control the inverter such that the actual inverter current will track the reference inverter current. Therefore, in order to achieve effective grid synchronization of a three-phase PV system, we first designed a grid integration control scheme in which Incremental Conductance (IC) is employed for the maximum power tracking control, Instantaneous Power Theory (IPT) for estimation of reference inverter current and a Sliding Mode Controller (SMC) for generating appropriate inverter switching signals.
The output of a PV system varies with variation in solar irradiance and temperature. Further, the V-I characteristics of a PV panel is nonlinear, therefore it is essential to operate a PV panel at the Maximum Power Point (MPP) in order to extract maximum power from it. This can be accomplished by designing an efficient Maximum Power Point Tracking (MPPT) algorithm for a PV system that can handle the situation of varied solar irradiance and temperature. But in order to handle the above situation, IC MPPT algorithm is found to be ineffective. Hence, a Reinforcement Learning (RL) based adaptive MPPT algorithm is proposed in this thesis to extract maximum output power from the PV panels during the changes in solar irradiance and temperature. The performance of the proposed RL MPPT algorithm is compared with that of IC MPPT algorithm and an advance MPPT algorithm e.g. Fuzzy Logic (FL) based MPPT algorithm. It is observed from the obtained results that the proposed RLMPPT algorithm exhibits improved maximum power tracking performance than both the IC MPPT and FL MPPT algorithms when there are variations in solar irradiance and temperature.
In order to achieve the grid power factor unity and the THD of the grid current should be within the permissible limits of IEEE-519 standard, a suitable estimation algorithm is essential for estimating the reference inverter current. Therefore, in developing the control schemes for grid synchronization of a PV system the estimation of reference current and its tracking are necessary. Some of the reported algorithms for estimation of reference currents e.g. Synchronous Reference Frame (SRF), Instantaneous Power Theory (IPT) have limitations in the situations of grid disturbances and load variations. The advanced algorithms e.g. Least mean square (LMS), Variable Step Size-Least Mean Square (VSS-LMS), Leaky Least Mean Square (LLMS) have also limitations such as slow convergence rate, parameter drifting and more complexity. It is thus necessary to develop an adaptive estimation algorithm for estimation of reference inverter current. In view of resolving the aforesaid limitations, we propose to employ the Variable Leaky Least Mean Square (VLLMS) algorithm for estimation of reference inverter current. The VLLMS algorithm is employed for the ideal grid (without grid voltage distortion), distorted grid, unbalanced grid, unbalanced and distributed grid, unbalanced load and variable solar irradiance conditions. The performance of the VLLMS algorithm is compared with that of the Leaky Least Mean square (LLMS) and Instantaneous Power Theory (IPT) based reference inverter current estimation algorithms. It is observed from the comparison that VLLMS algorithm outperforms in reducing the THD with no overshoot and undershoot in the grid current waveform during the aforesaid situations.
However, for generation of switching signals for inverter to track the reference inverter current, firstly we employ a sliding mode controller to handle the uncertain dynamics of the inverter. The performance of sliding mode controller is compared with the other two popular control algorithms e.g. hysteresis controller and adaptive hysteresis controller. Further, in order to handle the parametric variations in the VSI e.g. inductance and resistance of VSI and the variations in the load and grid disturbances, we design an Adaptive Sliding Mode Controller. The performance of adaptive sliding mode controller is compared to that of sliding mode controller during the parametric variation, load variation and grid disturbances.
All the proposed PV grid integration control schemes namely Incremental Conductance- Instantaneous Power Theory-Sliding Mode Controller (IC-IPT-SMC), Reinforcement Learning-IPT-SMC (RL-IPT-SMC), RL-Variable Leaky Least mean Square- SMC (RL-VLLMS-SMC) and RL-VLLMS-Adaptive Sliding Mode Controller (RLVLLMS- ASMC) were implemented in MATLAB R/Simulink Rand then in real-time on a PV prototype setup developed in the laboratory. The performances of all these controllers are evaluated and it is found that RL-VLLMS-ASMC exhibits superior performance in terms of effective maximum power extraction, fast and robust harmonic compensation during variation in solar irradiance and temperature, parametric uncertainties, grid disturbances and load variations.
|Item Type:||Thesis (PhD)|
|Uncontrolled Keywords:||Maximum power point tracking, Sliding mode control; Reinforcement learning; Instantaneous power theory; Variable Leaky least mean square; Leaky least mean square algorithm; Adaptive sliding mode control|
|Subjects:||Engineering and Technology > Electrical Engineering > Power Systems|
Engineering and Technology > Electrical Engineering > Power Networks
|Divisions:||Engineering and Technology > Department of Electrical Engineering|
|Deposited By:||IR Staff BPCL|
|Deposited On:||22 Feb 2019 20:48|
|Last Modified:||22 Feb 2019 20:48|
|Supervisor(s):||Subudhi, Bidyadhar and Ray, Pravat Kumar|
Repository Staff Only: item control page