This article introduces an innovative parallel differential power processing (PDPP) architecture designed to mitigate the effect of mismatch among photovoltaic (PV) strings. The proposed PV to virtual bus PDPP architecture leverages a virtual bus as the input for all string-level converters. Notably, this approach allows for a reduction in the voltage rating of components since the virtual bus voltage can be set lower than the main bus or PV strings voltage. In this architecture, crucial requirements for the string-level converters (SLCs) include the capability to generate positive and negative output voltages and to provide isolation. To fulfill these requirements, a dual active bridge converter connected to a bridgeless converter as the PDPP SLCs is considered. In this architecture, while SLCs ensure maximum power point tracking (MPPT) for each PV string using conventional MPPT algorithms, the central converter controls the virtual bus voltage. Experimental results validate the performance of the proposed PV to virtual bus PDPP architecture with a system efficiency ranging from 96.4% to 99%.

This thesis introduces and develops advanced Differential Power Processing (DPP) architectures to enhance the performance and efficiency of photovoltaic (PV) systems by addressing mismatches among PV modules and strings. The work focuses on mitigating power losses due to voltage and current mismatches due to factors such as partial shading, panel misalignment, dust, and degradation. The research proposes novel architectures designed to reduce component voltage and power ratings, potentially lowering costs while maintaining high efficiency.

PV to Virtual Bus Parallel Differential Power Processing (PV2VB PDPP) Architecture: A new PV2VB PDPP architecture is introduced, leveraging a virtual bus as the input for string-level converters (SLCs). This design allows for reduced components’ voltage ratings by operating the virtual bus at a lower voltage than the main bus or PV strings. The architecture employs Dual Active Bridge converters connected to Bridgeless converters as SLCs to provide isolation and handle both positive and negative outputs. Experimental results demonstrate system efficiency ranging from 96.4% to 99%.

Dynamic Analysis and Stability: The thesis includes a comprehensive dynamic analysis of the PV2VB PDPP architecture, deriving small-signal models, transfer functions, and frequency responses. These analyses aid in understanding the system’s dynamic behavior, enabling effective controller design and stability studies. Experimental validation confirms fast stabilization of the virtual bus voltage (0.6 seconds) and intermediate bus voltages (15 milliseconds), ensuring efficient Maximum Power Point Tracking (MPPT) for each PV string.

Battery Integration in PV2VB PDPP Architecture: The work extends the PDPP architecture to include battery integration at the virtual bus, facilitating energy storage and management while performing MPPT. The battery integration reduces component voltage ratings and allows for efficient charging and discharging control by the central converter. Experimental evaluations show system efficiencies between 95.5% and 99%.

PV to Virtual Bus Series-Parallel Differential Power Processing (PV2VB SPDPP) Architecture: To address mismatches in both series-connected modules and parallel-connected strings, a PV2VB SPDPP architecture is proposed. This architecture uses a combination of SLCs and module-integrated converters (MICs), processing only a fraction of the total power. By leveraging virtual buses for both SLCs and MICs, the architecture reduces voltage and power stress on components, improving cost-effectiveness and reliability. Real-time simulations validate the system’s ability to balance power flow, ensure stable operation, and optimize PV module performance under mismatch conditions.

Photovoltaic (PV) to virtual bus parallel differential power processing (PDPP) architecture can mitigate mismatch losses among PV strings. This article presents a comprehensive dynamic analysis by deriving a small-signal model of the PDPP architecture based on its state space model. Subsequently, the corresponding transfer functions and frequency response are obtained, offering valuable insights into the dynamic behavior of the architecture. To validate the accuracy of the derived model, the frequency response has also been achieved by observed data from both PLECS simulation and experiment through system identification. Besides, this article discusses the design considerations of the discrete controllers' parameters for both virtual and intermediate bus voltages and studies the stability of the architecture. Experimental measurements confirm the ability of the central controller to stabilize the virtual bus voltage to the desired level within 0.6 seconds, while the intermediate bus voltages settle within 15 ms, enabling proper maximum power point tracking of each PV string.

Photovoltaic (PV) systems are often exposed to mismatch caused by partial shading, different mounting angles, dust accumulation, cell degradation, and so on. This paper proposes a novel parallel differential power processing (P-DPP) configuration to minimize mismatch-related losses among PV strings. The proposed configuration, called PV to Virtual Bus P-DPP, uses a virtual bus as an input for all P-DPP converters. Since the virtual bus voltage can be selected lower than the DC Bus voltage, components’ voltage rating can be reduced. An essential feature of the proposed configuration is the ability of the converters to generate both positive and negative output voltage. Therefore, a bidirectional flyback converter connected to a bridgeless converter is proposed as the P-DPP converter. To find the MPP of each PV string, the Perturb and Observe (P&O) algorithm is implemented. Moreover, a proportional–integral feedback controller controls the virtual bus voltage through the central converter. The benefits of the proposed configuration are discussed, and the operation of the proposed structure is further verified through simulations with the software PLECS.