ABSTRACT:
Residential solar photovoltaic (PV) energy is becoming
an increasingly important part of the world's renewable energy. A residential
solar PV array is usually connected to the distribution grid through a
single-phase inverter. Control of the single-phase PV system should maximize the
power output from the PV array while ensuring overall system performance,
safety, reliability, and controllability for interface with the electricity
grid. This paper has two main objectives. The first objective is to develop an
artificial neural network (ANN) vector control strategy for a LCL-filter based single-phase
solar inverter. The ANN controller is trained to implement optimal control,
based on approximate dynamic programming. The second objective is to evaluate
the performance of the ANN-based solar PV system by (a) simulating the PV system
behavior for grid integration and maximum power extraction from solar PV array
in a realistic residential PV application and (b) building an experimental
solar PV system for hardware validation. The results demonstrate that a
residential PV system using the ANN control outperforms the PV system using the
conventional standard vector control method and proportional resonant control
method in both simulation and hardware implementation. This is also true in the
presence of noise, disturbance, distortion, and non-ideal conditions.
KEYWORDS:
1. Artificial neural networks
2. DC-AC power converters
3. DC-DC power converters
4. Dynamic programming
5. Maximum power point tracker
6. Optimal control
7. Solar power generation
SOFTWARE:
MATLAB/SIMULINK
CONCLUSION:
This
paper proposes a single-phase, residential solar PV system based on artificial
neural networks and adaptive dynamic programming for MPPT control and grid
integration of a solar photovoltaic array through an LCL-filter based inverter.
The proposed artificial neural network controller implements the optimal
control based on the approximate dynamic programming. Both the simulation and
hardware experiment results demonstrate that the solar PV system using the
ADP-based artificial neural network controller has more improved performance
than that using the proportional resonant or conventional standard vector
control techniques, such as no requirement for damping resistance, more
reliable and efficient extraction of solar power, more stable DC-link voltage,
and more reliable integration with the utility grid. Using the ADP-based neural
network control technique, the harmonics are significantly reduced and the
system shows much stronger adaptive ability under uncertain conditions, which
would greatly benefit the integration of small-scale residential solar
photovoltaic systems into the grid.
REFERENCES:
[1]
Renewable Energy World Editors. (2014, Nov. 12). Residential Solar Energy
Storage Market Could Approach 1 GW by 2018. Available:http://www.renewableenergyworld.com.
[2]
R. A. Mastromauro, M. Liserre and A. D. Aquila, “Control Issues in Single-Stage
Photovoltaic Systems: MPPT, Current and Voltage Control”, IEEE Trans. Ind.
Informatics, vol. 8, no. 2, pp. 241-254, May 2012.
[3]
E. Lorenzo, G. Araujo, A. Cuevas, M. Egido, J. Miñano and R. Zilles, Solar
Electricity: Engineering of Photovoltaic Systems, Progensa, Sevilla, Spain,
1994.
[4]
J. M. Carrasco, L. G. Franquelo, J. T. Bialasiewicz, E. Galván, R. C. P. Guisado,
M. Á. M. Prats, J. I. León, and N. Moreno-Alfonso, “Power- Electronic Systems
for the Grid Integration of Renewable Energy Sources: A Survey”, IEEE Trans.
Ind. Electron., vol. 53, no. 4, pp. 1002-1016, August 2006.
[5]
W. T. Franke, C. Kürtz and F. W. Fuchs, "Analysis of control strategies for
a 3 phase 4 wire topology for transformerless solar inverters," inProc.
IEEE Int. Symp. Ind. Electron., Bari, pp. 658-663, 2010.
