Control of the photovoltaic emulator using fuzzy logic based resistance feedback and binary search
Photovoltaic (PV) emulator is a power supply that produces similar currentvoltage (I-V) characteristics as the PV module. This device simplifies the testing phase of PV systems under various conditions. The essential part of the PV emulator (PVE) is the control strategy. Its main function is to dete...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2018
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/81690/1/RazmanAyopPSKE2018.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Photovoltaic (PV) emulator is a power supply that produces similar currentvoltage (I-V) characteristics as the PV module. This device simplifies the testing phase of PV systems under various conditions. The essential part of the PV emulator (PVE) is the control strategy. Its main function is to determine the operating point based on the load of the PVE. The direct referencing method (DRM) is the widely used control strategy due to its simplicity. However, the main drawback of DRM is that the output voltage and current oscillate due to the inconsistent operating point under fixed load. This thesis proposes an improved and robust control strategy named resistance feedback method (RFM) that yields consistent operating point under fixed load, irradiance and temperature. The RFM uses the measured voltage and current to determine the load of the PVE in order to identify the accurate operating point instantaneously. The conventional PV models include the I-V and voltage-current PV model. These PV models are widely used in various control strategies of PVE. Nonetheless, the RFM requires a modified PV model, the current-resistance (I-R) PV model, where the mathematical equation is not available. The implementation of the I-R PV model using the look-up table (LUT) is feasible, but it requires a lot of memory to store the data. A mathematical equation based I-R PV model computed using the binary search method is proposed to overcome the drawback of the LUT. The RFM consists of the I-R PV model and the closed-loop buck converter. In this work, the RFM is investigated with two different controllers, namely the proportional-integral (PI) and fuzzy logic controllers. The RFM using the PI controller (RFMPI) and the RFM using the fuzzy logic controller (RFMF) are tested with resistive load and maximum power point tracking (MPPT) boost converter. The perturb and observe algorithm is selected for the MPPT boost converter. In order to properly design the boost converter for the MPPT application, the sizing of the passive components is proposed, derived and confirmed through simulation. This derivation allows adjustment on the output voltage and current ripple of the PVE when connected to the MPPT boost converter. The simulation results of the proposed control strategies are benchmarked with the conventional DRM. To validate the simulation results, all controllers are implemented using dSPACE ds1104 rapid prototyping hardware platform. The RFM computes an operating point of the PVE at 20% faster than the DRM. The generated output PVE voltage and current using RFMPI and the RFMF are up to 90% more accurate compared to the DRM. The efficiency of the PVE is beyond 90% when tested under locus of maximum power point. In transient analysis, the settling time of RFMF is faster than the RFMPI. In short, the proposed RFMF is robust, accurate, quick respond and compatible with the MPPT boost converter. |
|---|