Integração energética de biorrefinarias de cana-de-açúcar com operação multiperiódica

Abstract

Brazil has a great importance as biofuels producer, especially of ethanol from sugarcane. The bagasse, a byproduct of sugarcane biorefineries, can be used to generate electricity and produce second generation ethanol. However, variations in prices of electricity and ethanol may motivate variations in the operating conditions of the integrated biorefinery process, which produces first and second generation ethanol and bioelectricity. The heat exchanger network (HEN) of such a process must be able to meet these variations. This work aimed the synthesis of multiperiod HEN in sugarcane biorefineries. In this approach, each period indicates a process condition and the HEN synthesized is able to meet these different operating conditions. Three industrial case studies are considered. Case Study 1 (CS1) is a biorefinery that produces 1G/2G ethanol and electricity, disposing the pentoses fraction. In Case Studies 2 and 3, CS2 and CS3, the process is similar, but the pentoses fraction is used to produce ethanol (CS2) or biogas (CS3). For each biorefinery, three periods were considered, which differ in the bagasse fraction diverted to 2G ethanol production. In each period, a Mixed Integer Nonlinear Programming (MINLP) problem was solved to minimize the total annualized cost (TAC) and timesharing mechanisms were used to integrate the HENs of all periods into a multiperiod HEN. Optimization problems were solved at two levels using three different strategies. In the first strategy, an adapted Particle Swarm Optimization algorithm was used in both levels. However, the solutions by this method presented small TAC improvements compared to the process commonly found in Brazilian plants, where there is already energy integration among some process streams (called in this work of the process with project integration). To deal with this problem, two strategies were employed with hybrid metaheuristics: Simulated Annealing and Rocket Fireworks Optimization; and Tabu Search and Particle Swarm Optimization. For processes with the multiperiod HENs, these latter two strategies presented reductions above 58% and 54% in TAC and steam demand, respectively, compared to processes without energy integration. Also, using the aforementioned methods, improvements in TAC and steam demand for process with the multiperiod HENs reach values above 44% and 41%, respectively, in relation to processes with project integration. Such reductions in steam demand allow diverting more bagasse to produce second generation ethanol. In addition, energy integration in biorefineries provides improved energy management and reduced operating and capital costs. Thus, all these improvements contribute to 1G/2G ethanol and electricity production process.