Overview of the Grid-Connected PV System
A grid-connected photovoltaic system consists of a solar PV panel array that is connected to the main power grid. The system includes a transformer, AC loads, and an inverter that converts the DC power from the solar panels into AC power, which is supplied to the grid. The focus here is on integrating MPPT with the grid-connected system to ensure the maximum power is extracted from the solar PV panels, regardless of varying sunlight conditions.
System Configuration:
Transformer: 34 kV to 400 V step-up transformer.
AC Loads: Two connected loads, with 500 kW and 30 kW ratings.
PV System: The PV array is rated at 41 kW with a line-to-line voltage of 400 V, providing a clean DC supply to the grid.
Specifications of the Solar PV Panel
The PV panels used in this system have the following characteristics:
Power per Panel: 414 W.
Open Circuit Voltage (Voc): 85.3 V.
Maximum Power Voltage (Vmp): 72.9 V.
Short Circuit Current (Isc): 6.09 A.
Maximum Power Current (Imp): 5.69 A.
The PV system operates efficiently under standard irradiance of 1000 W/m², and the output power varies according to the irradiance level. At full irradiation, the system generates up to 40 kW, while under reduced irradiance, the output power decreases accordingly.
The Role of Artificial Neural Networks in MPPT
MPPT is a technique used to maximize the power extracted from solar panels by continuously adjusting the operating point of the panels. In this system, an Artificial Neural Network (ANN) is employed to predict the optimal operating point based on real-time measurements of temperature and irradiation. The ANN model helps generate a reference voltage that is fed into the inverter to control power flow.
Key Components:
Input Data: Irradiation and temperature data are used as inputs for training the neural network.
Output Data: The network generates the reference maximum power voltage, which helps in controlling the inverter's operation.
The ANN is trained using a dataset that includes varying temperature and irradiation conditions, ensuring that the system can adapt to changing environmental factors.
PV Model and Data Collection for ANN Training
To train the neural network, we begin by defining the PV panel's electrical characteristics using standard equations for maximum current, voltage, and power. The temperature and irradiation conditions are randomly varied within predefined ranges:
Temperature Range: 15°C to 35°C.
Irradiation Range: 0 to 1000 W/m².
This data is used to calculate the corresponding power output from the solar panels, which is then used to generate the reference voltage required for the MPPT algorithm. The neural network is trained using this input-output data, with high correlation (R-value = 1) ensuring that the network can accurately predict the reference voltage.
Control System for Power Flow Management
Once the neural network is trained, it is integrated into the system’s control logic. The ANN model provides the reference voltage, which is sent to a voltage regulator. The regulator then controls the inverter to extract the maximum possible power from the PV system.
Reference Current Generation: Based on the reference voltage, the current controller generates a reference for the direct-axis current (Id) to ensure that only real power is sent from the PV system to the grid.
Reactive Power Management: The system avoids reactive power flow, ensuring that only active power is supplied to the grid.
The system constantly monitors and adjusts the power flow, ensuring that the grid receives power efficiently while optimizing the performance of the solar panels.
Simulation and Testing with Varying Load Conditions
The system is simulated under different load and irradiation conditions to verify its performance.
Constant Irradiation: When irradiation is set to 1000 W/m², the system generates approximately 40.6 kW, with any excess power being sent to the grid.
Variable Load Conditions: When the load changes, the system dynamically adjusts, ensuring that the required power is supplied to both the load and the grid. For instance, if the load reduces to 30 kW, the excess power is fed into the grid.
The control system ensures that the grid is balanced at all times, adapting to the changes in load and solar power generation.
Impact of Changing Irradiation on System Performance
In real-world conditions, the irradiation levels fluctuate, affecting the power output from the PV system. To test how the system reacts to these changes, the irradiation is reduced from 1000 W/m² to 500 W/m². As expected, the power generated by the PV system drops from 40.6 kW to 20.2 kW. During this time, the system draws additional power from the grid to meet the load demand.
This adaptive nature of the system ensures that the grid load is always met, while the MPPT algorithm continues to extract the maximum power possible from the PV panels.
Conclusion
The integration of an Artificial Neural Network-based MPPT system in a grid-connected PV system proves to be highly effective in maximizing solar power output and ensuring efficient power management. The neural network model accurately predicts the reference voltage needed for optimal power extraction, and the system adapts well to varying irradiation and load conditions. By continuously adjusting the power flow from the solar panels to the grid, this system contributes to a more efficient and reliable renewable energy source.
Kommentarer