Estudo de curvas de penalidade para controle de processos aplicado à cogeração de uma planta sucroalcooleira

• Autor(a) principal: Torres, Matheus Henrique Cândido
• Data de Publicação: 2025
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Institucional da UFSCAR
• Texto Completo: https://hdl.handle.net/20.500.14289/21880
• topic: Controle de processos; Função penalidade; Otimização de processos; Cogeração; Turbogerador; Geração de vapor; Process control; Penalty function; Process optimization; Cogeneration; Turbo generator; Steam generation; ENGENHARIAS::ENGENHARIA QUIMICA::PROCESSOS INDUSTRIAIS DE ENGENHARIA QUIMICA

In cogeneration systems of sugarcane processing plants, steam is generated in boilers from the combustion heat of sugarcane bagasse to meet variable demands of mechanical drives and heating in process equipment, while simultaneously producing electric power through steam expansion in turbo generators. The cogeneration system is subject to various disturbances caused by fluctuations in utility demands or in the high-pressure steam supply from the boiler. Under such circumstances, critical demands for factory operation must be prioritized with maximum energy efficiency. In many cases, a supervisory system is already in place to suggest adjustments to the flow rates of various steam lines in order to maintain process steam supply, commonly through PID controllers, a single-input and single-output system used to control a multivariable process. In this context, a periodic tuning method for controllers was proposed, comparing proportional and proportional-integral controllers in a cogeneration system, with the objective of improving stability in the steam lines and maximizing energy efficiency. Two optimization algorithms were tested for tuning the proportional and integral gains, BFGS and SLSQP, applied to an objective function that evaluates the sum of the deviation from the set point in steam line pressures and the surplus power generation. The best performance was achieved with the SLSQP method, which imposed constraints on the maximum and minimum values of the proportional and integral gains, considering a 40-second optimization horizon. Two penalty functions for exported energy, quadratic and linear, were tested, with the quadratic function resulting in greater electric energy export but with a higher deviation from the pressure set points. Overall, the control system demonstrated robustness in correcting disturbances and optimizing electric power generation.