Please use this identifier to cite or link to this item: http://dspace.dtu.ac.in:8080/jspui/handle/repository/19308
Full metadata record
DC FieldValueLanguage
dc.contributor.authorPATHANIA, VIVEK-
dc.date.accessioned2022-07-28T10:13:18Z-
dc.date.available2022-07-28T10:13:18Z-
dc.date.issued2022-05-
dc.identifier.urihttp://dspace.dtu.ac.in:8080/jspui/handle/repository/19308-
dc.description.abstractEnergy consumption has been steadily growing and is expected to persist in the future. This increasing demand for energy encourages the fast growth of renewable energy sources for instance solar, wind, tidal, geothermal, and others in order to reduce fossil fuel usage and keep environment pollution free. Aside from wind power, solar energy is the most widely utilised energy source, having a large market share in the global energy business. In recent times with significant developments in the field of photovoltaic (PV) panels the costs for PV systems have also been reduced. With advancements in power electronics converters for better system performance, an efficient maximum power point tracking (MPPT) controller is required to increase system throughput. To ensure that the maximum power point (MPP) can always meet the objective under varying weather circumstances of solar radiation and temperature, the MPPT algorithm is used in conjunction with a DC/DC converter. Along with a bidirectional converter (BDC) coupled with the battery system for utilising the stored power in times of less or no solar irradiance. This solar converter typically consists of DC-DC boost converter as the MPP converter and a bidirectional buck-boost converter as BDC. In modern times with the availability of faster computing devices and developments in artificial intelligence (AI), new and more adaptable MPPT controllers can be designed around AI. One such method is deep reinforcement learning which uses deep neural networks to model out the state-action in any given environment. Using a deep deterministic policy gradient algorithm in MPPT ensures a continuous action space available for controlling the MPP converter and BDC. The MPP converter is responsible for achieving the maximum operation condition for solar modules whereas, the BDC is accountable for maintaining a steady DC bus voltage through charging and discharging of battery system. Loads such as DC loads and AC loads with a suitable inverter can be coupled to this DC bus to be utilised in household applications or electric vehicle charging stations.en_US
dc.language.isoenen_US
dc.relation.ispartofseriesTD-5863;-
dc.subjectSOLAR POWER MANAGEMENen_US
dc.subjectDEEP REINFORCEMENT LEARNINGen_US
dc.subjectPV PANELSen_US
dc.subjectMPPT ALGORITHMen_US
dc.subjectDC-DC CONVERTERen_US
dc.titleDEVELOPMENT OF SOLAR POWER MANAGEMENT SYSTEM USING DEEP REINFORCEMENT LEARNINGen_US
dc.typeThesisen_US
Appears in Collections:M.E./M.Tech. Electrical Engineering

Files in This Item:
File Description SizeFormat 
VIVEK PATHANIA M.Tech.pdf2.84 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.