{"id":{"repo_id":"southwales","oai_identifier":"oai:pure.atira.dk:studenttheses/806749d1-0119-40d1-b304-7e8f26c66aed"},"canonical_url":"https://search.dev.ndltd.org/etd/southwales/oai:pure.atira.dk:studenttheses/806749d1-0119-40d1-b304-7e8f26c66aed","repository":{"repo_id":"southwales","name":"University of South Wales","base_url":"https://pure.southwales.ac.uk/ws/oai"},"display":{"title":"Research on Dual-Axis Solar Tracking Systems Using AI-Controlled Boost Converter for Sustainable Energy Solutions","abstract":"Solar photovoltaic (PV) systems lose significant energy yield when panels cannot track the sun dynamically. Dual-axis tracking systems address this limitation but require precise orientation estimation and efficient power conditioning, both of which remain technically challenging in embedded and simulation-driven contexts. <br/><br/>There are two interrelated contributions in the dissertation to intelligent solar energy solutions. Firstly, an augmented PID controller based on the Kalman filter is implemented to control orientation along the elevation axis of a dual-axis tracker with the use of a microprocessor. Two states' Kalman filter is utilized in combination with data received from an accelerometer and gyroscope of the LSM6DS3TRC sensor. The estimated drift-compensated angles actuate a PID controller which drives a DC permanent magnet motor. It is compared to a second-order LPF IIR filter used in identical experimental setup. Secondly, a five-layer hierarchical structure with three AI modules is constructed to analyse, predict, and control a DC-DC boost converter. The data used for training purposes were obtained via Simulink/MATLAB simulation with parameters corresponding to actual MPPT telemetry. <br/><br/>The Kalman filter reduced steady-state error and RMSE by 43.59% and 28.57%, respectively. An accuracy of prediction in power calculation was found to be R² ≈ 0.989, while MAE for estimation of efficiency was only 0.0043. Fault detection achieved perfect F1= 1.0, and macro-F1 for supervised control of the duty cycle was 0.97. Finally, the reinforcement learning model yielded an agent with an average reward of about 558. Therefore, the research shows how Kalman filter can be used for state estimation and how AI techniques improve traditional energy systems.","abstract_html":"Solar photovoltaic (PV) systems lose significant energy yield when panels cannot track the sun dynamically. Dual-axis tracking systems address this limitation but require precise orientation estimation and efficient power conditioning, both of which remain technically challenging in embedded and simulation-driven contexts. &lt;br/&gt;&lt;br/&gt;There are two interrelated contributions in the dissertation to intelligent solar energy solutions. Firstly, an augmented PID controller based on the Kalman filter is implemented to control orientation along the elevation axis of a dual-axis tracker with the use of a microprocessor. Two states&#x27; Kalman filter is utilized in combination with data received from an accelerometer and gyroscope of the LSM6DS3TRC sensor. The estimated drift-compensated angles actuate a PID controller which drives a DC permanent magnet motor. It is compared to a second-order LPF IIR filter used in identical experimental setup. Secondly, a five-layer hierarchical structure with three AI modules is constructed to analyse, predict, and control a DC-DC boost converter. The data used for training purposes were obtained via Simulink/MATLAB simulation with parameters corresponding to actual MPPT telemetry. &lt;br/&gt;&lt;br/&gt;The Kalman filter reduced steady-state error and RMSE by 43.59% and 28.57%, respectively. An accuracy of prediction in power calculation was found to be R² ≈ 0.989, while MAE for estimation of efficiency was only 0.0043. Fault detection achieved perfect F1= 1.0, and macro-F1 for supervised control of the duty cycle was 0.97. Finally, the reinforcement learning model yielded an agent with an average reward of about 558. Therefore, the research shows how Kalman filter can be used for state estimation and how AI techniques improve traditional energy systems.","abstract_has_math":false,"creators":["Bardhan Promi, Nirbachita"],"institution":null,"degree_name":"Master's Thesis","degree_level":"Student thesis","degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Nazir, Mian Hammad","Bai, Jiping","Sivanathan, Sivagunalan"],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026","date_published":"2026","updated_at":"2026-07-24T04:40:03Z","subjects":[],"languages":["eng"],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["oai:pure.atira.dk:studenttheses/806749d1-0119-40d1-b304-7e8f26c66aed"],"render_values":[{"text":"oai:pure.atira.dk:studenttheses/806749d1-0119-40d1-b304-7e8f26c66aed","href":null,"code":true}]}]},"links":{"outbound_url":"https://pure.southwales.ac.uk/en/studentTheses/806749d1-0119-40d1-b304-7e8f26c66aed","outbound_label":"Repository record","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Nazir, Mian Hammad","Bai, Jiping","Sivanathan, Sivagunalan"]},{"key":"dc:creator","label":"Author","values":["Bardhan Promi, Nirbachita"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2026"]},{"key":"dc:date.issued","label":"Date","values":["2026"]},{"key":"dc:relation.isreferencedby","label":"Dc Relation Isreferencedby","values":["https://pure.southwales.ac.uk/en/studentTheses/806749d1-0119-40d1-b304-7e8f26c66aed"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"dc:type.qualificationlevel","label":"Dc Type Qualificationlevel","values":["Student thesis"]},{"key":"dc:type.qualificationname","label":"Dc Type Qualificationname","values":["Master's Thesis"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["oai:pure.atira.dk:studenttheses/806749d1-0119-40d1-b304-7e8f26c66aed","https://pure.southwales.ac.uk/en/studentTheses/806749d1-0119-40d1-b304-7e8f26c66aed"]},{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://pure.southwales.ac.uk/files/37884015/Nirbachita_Bardhan_Promi_MRES_Final_id.pdf"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Solar photovoltaic (PV) systems lose significant energy yield when panels cannot track the sun dynamically. Dual-axis tracking systems address this limitation but require precise orientation estimation and efficient power conditioning, both of which remain technically challenging in embedded and simulation-driven contexts. <br/><br/>There are two interrelated contributions in the dissertation to intelligent solar energy solutions. Firstly, an augmented PID controller based on the Kalman filter is implemented to control orientation along the elevation axis of a dual-axis tracker with the use of a microprocessor. Two states' Kalman filter is utilized in combination with data received from an accelerometer and gyroscope of the LSM6DS3TRC sensor. The estimated drift-compensated angles actuate a PID controller which drives a DC permanent magnet motor. It is compared to a second-order LPF IIR filter used in identical experimental setup. Secondly, a five-layer hierarchical structure with three AI modules is constructed to analyse, predict, and control a DC-DC boost converter. The data used for training purposes were obtained via Simulink/MATLAB simulation with parameters corresponding to actual MPPT telemetry. <br/><br/>The Kalman filter reduced steady-state error and RMSE by 43.59% and 28.57%, respectively. An accuracy of prediction in power calculation was found to be R² ≈ 0.989, while MAE for estimation of efficiency was only 0.0043. Fault detection achieved perfect F1= 1.0, and macro-F1 for supervised control of the duty cycle was 0.97. Finally, the reinforcement learning model yielded an agent with an average reward of about 558. Therefore, the research shows how Kalman filter can be used for state estimation and how AI techniques improve traditional energy systems."]},{"key":"dc:title","label":"Title","values":["Research on Dual-Axis Solar Tracking Systems Using AI-Controlled Boost Converter for Sustainable Energy Solutions"]}]}],"canonical_facts":{"dc:contributor.advisor":["Nazir, Mian Hammad","Bai, Jiping","Sivanathan, Sivagunalan"],"dc:creator":["Bardhan Promi, Nirbachita"],"dc:date":["2026"],"dc:date.issued":["2026"],"dc:description.abstract":["Solar photovoltaic (PV) systems lose significant energy yield when panels cannot track the sun dynamically. Dual-axis tracking systems address this limitation but require precise orientation estimation and efficient power conditioning, both of which remain technically challenging in embedded and simulation-driven contexts. <br/><br/>There are two interrelated contributions in the dissertation to intelligent solar energy solutions. Firstly, an augmented PID controller based on the Kalman filter is implemented to control orientation along the elevation axis of a dual-axis tracker with the use of a microprocessor. Two states' Kalman filter is utilized in combination with data received from an accelerometer and gyroscope of the LSM6DS3TRC sensor. The estimated drift-compensated angles actuate a PID controller which drives a DC permanent magnet motor. It is compared to a second-order LPF IIR filter used in identical experimental setup. Secondly, a five-layer hierarchical structure with three AI modules is constructed to analyse, predict, and control a DC-DC boost converter. The data used for training purposes were obtained via Simulink/MATLAB simulation with parameters corresponding to actual MPPT telemetry. <br/><br/>The Kalman filter reduced steady-state error and RMSE by 43.59% and 28.57%, respectively. An accuracy of prediction in power calculation was found to be R² ≈ 0.989, while MAE for estimation of efficiency was only 0.0043. Fault detection achieved perfect F1= 1.0, and macro-F1 for supervised control of the duty cycle was 0.97. Finally, the reinforcement learning model yielded an agent with an average reward of about 558. Therefore, the research shows how Kalman filter can be used for state estimation and how AI techniques improve traditional energy systems."],"dc:identifier":["oai:pure.atira.dk:studenttheses/806749d1-0119-40d1-b304-7e8f26c66aed","https://pure.southwales.ac.uk/en/studentTheses/806749d1-0119-40d1-b304-7e8f26c66aed"],"dc:identifier.uri":["https://pure.southwales.ac.uk/files/37884015/Nirbachita_Bardhan_Promi_MRES_Final_id.pdf"],"dc:language":["eng"],"dc:relation.isreferencedby":["https://pure.southwales.ac.uk/en/studentTheses/806749d1-0119-40d1-b304-7e8f26c66aed"],"dc:title":["Research on Dual-Axis Solar Tracking Systems Using AI-Controlled Boost Converter for Sustainable Energy Solutions"],"dc:type":["Thesis"],"dc:type.qualificationlevel":["Student thesis"],"dc:type.qualificationname":["Master's Thesis"]},"updated_at":"2026-07-24T04:40:03Z"}