Modeling, simulation and control of hybrid power systems for vessels.

• Ano de defesa: 2022
• Autor(a) principal: Peixoto, Crisley de Souza
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Baterias; Controle Multinível; Eficiência energética; Embarcações marítimas; Energy efficiency and emissions; Hybrid power systems; Íons; Lithium-ion batteries; Lítio; Marine vessels; Multi-level control; Sistemas híbridos de potência
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/3/3150/tde-30112022-081832/

Climate concerns, regulations, and fuel costs are driving higher efficiency designs in the shipping industry. Research works suggest that hybrid power systems can be a viable option to reduce fuel consumption and emissions. However, more studies are required to analyze its net gains in different applications. In this work, we perform numerical simulations of powertrains for a diesel-electric Platform Supply Vessel (PSV) in a causal approach using MATLAB. To study such systems, we developed dynamical and static models for the power source representation and optimization, including models of diesel generator sets (gensets), a lithium-ion battery, and electric motors integrated with drivetrains and models that represent the vessel physics. All these models were combined using a static AC electrical network. The genset model was validated with errors below 2% for fuel consumption, and the engine speed results were kept within acceptable ranges, considering measured data from a project partner from the shipping industry. The battery cell model has shown errors below 1% for system states compared to literature results. We performed simulations considering hybrid and non-hybrid topologies, static models for optimization, and static and dynamical models for power source representation, comparing different energy management strategies. Also, we developed a representative response surface to model fuel consumption reductions using the software R to help in the hybrid power system design and selection of operation parameters. Through simulation experiments, it was possible to reduce fuel consumption, emissions, and genset running hours through strategic loading and hybridization using real operating profiles or a profile based on actual operations, considering different operational parameters. Through optimization, we observed the following reductions for the non-hybrid configuration: 5%-8% for fuel, 7%-8% for CO2, 23%-30% for genset operating time, and 27%-32% for particulate matter (PM). The NOx emissions varied, decreasing in 10% or increasing in 10%; however, an after-treatment system was not considered in any simulation. The genset redundancy had a negative and significant impact on all quantities of interest. The introduction of a battery brought additional benefits to the vessels power system compared to the non-hybrid topology, except for the NOx, which depended on the case. We compared the optimized non-hybrid operation with a hybrid configuration testing different energy management strategies, including ECMS formulations and a rule-based implementation. The comparisons indicate additional reductions that vary on the power profile but can be put into the ranges: 5%-10% for fuel, 5%-10% for CO2, 23%-49% for genset operating time, 26%-47% for PM. The NOx emissions decreased up to 11% and increased up to 9% depending on the simulation case. Most battery benefits were achieved in Dynamic Positioning (DP) or operating near the port. Energy efficiency increases above 3% were observed depending on the operating profile. Simulations that focused on the dynamic behaviour of the system showed that a 10% increase in the power system efficiency can be achieved by using a hybrid solution in a vessel typical acceleration ramp. The dynamic simulations also allowed the assessment of battery degradation, and the results indicated that a battery bank can last for around eight years of uninterrupted use, for a typical operating profile of a PSV.