Grid Connected Solar PV with Battery Energy Storage and ANN MPPT in MATLAB
 
This project demonstrates the modeling and simulation of a grid-connected photovoltaic energy system integrated with battery energy storage and an Artificial Neural Network (ANN)-based Maximum Power Point Tracking controller. The ANN MPPT receives normalized PV voltage and PV current as inputs and generates the duty cycle command for the PV boost converter.
The PV source supplies a common DC bus through a boost converter. A battery energy storage system is connected through a bidirectional DC-DC converter to support DC-bus voltage regulation and handle charging/discharging according to the power balance. A single-phase full-bridge inverter connects the DC system to the utility grid. Grid synchronization, current-reference generation, current regulation, and PWM generation are included in the inverter-control section.
 
Main System Features
 
Solar PV array modeling
ANN-based MPPT
ANN inputs: PV voltage Vpv​ and PV current Ipv​, in per-unit
ANN output: boost-converter duty cycle
PV boost DC-DC converter
Approximately 400 V DC bus
DC load connected to the common DC link
Battery Energy Storage System
Battery voltage approximately 260 V
Bidirectional buck-boost DC-DC converter
DC-bus voltage controller
Single-phase full-bridge grid-connected inverter
Grid synchronization/PLL
Current-reference generation
Current PI controller
PWM switching control
PV, battery, DC-load, DC-bus, inverter and grid measurements
ANN MPPT Dataset
 
The included ANN MPPT dataset contains 6,000 operating samples.
 
Parameter
Approximate Range
Solar irradiance
200-1000 W/m²
Temperature
15-55 °C
PV voltage
112-290 V
PV current
0.79-8.88 A
MPP voltage
193-244 V
MPP current
1.57-8.05 A
MPP power
317-1896 W
MPP duty ratio
0.390-0.517
ANN target duty
0.195-0.593
 
The dataset also contains normalized Vpv-pu and Ipv-pu variables for ANN training.
 
ANN Training Results
 
The package includes ANN training and validation plots. The recorded training results show:
 
Best validation MSE: 0.0015071
Best validation epoch: 8
Training regression: R = 0.89068
Validation regression: R = 0.88636
Testing regression: R = 0.89337
Overall regression: R = 0.89045
 
Simulation Operating Results
 
The simulation results demonstrate variable PV generation and corresponding battery power balancing.
 
The PV power changes approximately through:

2 kW → 1 kW → near 0 W → 1 kW → 2 kW
while the DC load remains approximately 1 kW. The bidirectional battery converter responds by alternating between charging and discharging. The DC-bus controller maintains the common bus close to 400 V despite changes in PV generation.
 
The battery voltage remains approximately 255-260 V, while battery current changes polarity according to power availability. The SOC plot confirms both charging and discharging operation.
The package also demonstrates grid-connected inverter voltage/current waveforms and DC-bus current balance.
 
Included Results
 
The uploaded package currently contains:
 
Complete Simulink system 
10-page PDF showing the Simulink architecture and controller subsystems
ANN MPPT dataset - 6,000 samples
ANN MPPT prediction/result dataset
ANN dataset 3-D map
Dataset sample plots
ANN training-performance plot
Training/validation/testing regression plots
Error histogram
ANN target-versus-prediction plots
Prediction-error plot
Test-data comparison
ANN architecture figure
PV voltage/current/power results
Battery voltage/current results
Battery SOC result
DC-bus voltage/load-current/load-power results
DC-bus current-balancing results
Inverter and grid voltage/current results
 
Applications
 
This model is suitable for:
 
B.Tech/B.E. projects
M.Tech/M.E. projects
Ph.D. research demonstrations
Renewable-energy laboratory studies
ANN-based MPPT research
Grid-connected PV studies
Battery energy-management studies
Intelligent power-electronics control research
Research-paper simulation validation