A Novel Distributed PV System Design

However, traditional PV configurations experience energy loss due to factors such as partial shading, panel mismatching, and temperature variation between panels. As a result, the PV system’s energy harvest efficiency is much less than expected.
We proposed a novel star-shaped distributed PV system that uses an FPGA embedded controller for parallel processing and voltage regulation for the PV panels. Another real-time controller is used for the operation of the MPPT to generate maximum energy from the PV panels. From simulation to prototype testing, including the design of the DC-to-DC converter for the distributed PV system, the NI graphical system design platform was used exclusively for this project. This work will lead to a state-of-the-art design of the PV power station, which will possibly be demonstrated as a building integrated photovoltaic public work for the Low-Carbon Emission Campus Project at National Cheng Kung University.

The Star-Shaped Distributed PV System
Each solar panel in the distributed PV system is equipped with a newly designed DC-to-DC converter, and the panel’s voltage output is regulated by an FPGA controller using proportional integral derivative control. Each solar panel’s power is optimized by another controller where the quadratic maximization MPPT algorithm is used to ensure that the panel’s output power is always maximized. The name “star shaped” is used because the distributed PV system is in cluster format, where a central console performs the control and system monitoring, and solar panels are connected to the central console. Monitoring the power generated by each of the clusters of panels in the entire system can be achieved through Ethernet or RS485 communication between central consoles.

PV Simulation and Converter Design
In the past, we used PSIM electrical software by Powersim to simulate PV power generation. We could then write a C code MPPT algorithm in C-Block, or call PSIM within LabVIEW via the Command Line Interface VI. The entire process of PSIM-LabVIEW simulation was very time-consuming because of call function between the two types of software. Now, with the Multisim and LabVIEW co-simulation process, we can write our QM MPPT algorithm in LabVIEW by directly calling our solar PV model established in Multisim. As a result, the simulation time decreased drastically, compared to previous work. Moreover, we can calibrate parameters used in the QM MPPT method and write the program to an NI embedded controller without extra work.

The design of the DC-to-DC converter used by the distributed PV system was also implemented using Multisim. The design work was then applied to a PCB layout via Ultiboard. Finally, the prototype of the converter was fabricated and tested for system verification. Designs for high-voltage input (for a thin-film solar panel) and a high-current input buck-boost converter (a polysilicon solar panel) were carried out successfully using this methodology.

The Distributed MPPT Controller
For the star-shaped distributed PV system design, a high-performance embedded controller was required to operate the voltage regulation and to run the MPPT algorithm for every PV module. We chose the NI sbRIO-9642XT embedded control and acquisition device, which has two embedded cores on the same platform. The FPGA chip is used for the parallel processing of the panel-level VI regulation, which includes signal acquisition, filtering, averaging, and closed-loop PID control strategy. This information is subsequently used by another real-time processor for the MPPT calculation.

Because of the graphical programming characteristics, the FPGA syntax under the LabVIEW environment was easy and straightforward. Only the serial peripheral interface protocol between the analog-to-digital converter and the synchronized PWM generator needed to be written by an engineer. The PID controller and digital filter can be created using the VI library, which saved a lot of development time for system implementation. With the LabVIEW shared variable function, we can monitor the power output for each of the PV panels. Even with portable devices such as the iPad, dashboard software can read information directly using the shared variable.

Experiment Results
A benchmark experiment was performed in two independent PV fields, one with a distributed PV system and the other one with the traditional centralized one. Four thin-film solar panels were connected in a series, and one of them was partially shaded. Daily power outputs for the four PV panels of the distributed system were 382 WH, 386 WH, 377 WH, and 261 WH. Test results showed that the proposed distributed systems worked well under panel-mismatching and partial-shading conditions. When compared with the total power output (1,192 WH, given by the centralized system), an 18 percent power increase was achieved by the proposed system.

Advantages of the NI Solution
Development of a PV power generation system required simulation using various design tools, including circuit design, MPPT simulation, DC-to-DC converter design, PV characteristic simulation, and system implementation of the embedded controller. If software and hardware from different vendors were used during development, it would make final system integration difficult and time-consuming. Seamless design tools provided by National Instruments helped us save a lot of time—from system design to prototype to integration to testing.

Future Application Development
The proposed distributed PV technology can offer many diversified designs for building an integrated PV application. For example, the research team for the Low-Carbon Emission Campus Project at National Cheng Kung University currently provides a creative, state-of-the-art design of a curved-surface scooter shelter. This is a green energy design that takes advantage of the novel star-shaped distributed PV technology. Power loss due to panel mismatching is no longer a problem, and energy efficiency increase is also expected.

Author(s):
Ru-Min Chao - National Cheng Kung University