• Energy Research

A system that combines a vapor compression refrigeration system (VCRS) with a vapor absorption refrigeration system (VARS) merges the advantages of both processes, resulting in a more cost-effective system. In such a cascade system, the electrical power for VCRS and the heat energy for VARS can be significantly reduced, resulting in a coefficient of performance (COP) value higher than the value of each system operating in standalone mode. A previously developed optimization model of a series flow double-effect H2O-LiBr VARS is extended to a superstructure-based optimization model to embed several possible configurations. This model is coupled to an R134a VCRS model. The problem consists in finding the optimal configuration of the cascade system and the sizes and operating conditions of all system components that minimize the total heat transfer area of the system, while satisfying given design specifications (evaporator temperature and refrigeration capacity of −17.0 °C and 50.0 kW, respectively), and using steam at 130 °C, by applying mathematical programming methods. The obtained configuration is different from those reported for combinations of double-effect H2O-LiBr VAR and VCR systems. The obtained optimal configuration is compared to the available data. The obtained total heat transfer area is around 7.3% smaller than that of the reference case.

Real-time optimization (RTO) is widely used in industry to operate processes close to their maximum performance. The models used for RTO need to be adapted using real-time data to ensure feasibility of the model optimal inputs and convergence to the real plant optimal point. Heat and power systems are suitable for being optimized in real-time because of their fast dynamics and the benefits achievable by reacting to changes in power prices and steam demand. This work proposes a modifier adaptation strategy that exploits the structure of certain problems to make the adaptation faster and more reliable, which is proven to be particularly useful for heat and power systems. The adaptation is performed in the equations that predict efficiencies or performance of unit operations. By identifying the variables that modify each performance factor, the number of data sets needed for gradient correction is reduced. This makes the proposed strategy suitable for real-time optimization of processes with a large number o...

Abstract This paper addresses the optimization of a single effect absorption refrigeration system operating with lithium bromide-water solution. A non-linear programming mathematical model is developed to determine the operating conditions and the distribution of the total heat transfer area (sizes) along the involved process units to optimize the following two objective functions: (i) maximization of the coefficient of performance for a given amount of the total heat transfer area, and (ii) minimization of the total heat transfer area of the system for a given cooling capacity. The proposed model can either be used for simulation or optimization purposes. Simulated or optimized values of temperature, pressure, composition and flow rate of all streams and sizing of each process unit are predicted. In addition, because of the non linear nature of the resulting model, a systematic solution procedure is proposed in order to guarantee the model convergence. A detailed discussion of the optimization results are presented through different case studies.

A nitrogen removal benchmark was analyzed using the Activated Sludge Models No. 1 (ASM1) and No. 3 (ASM3) in order to establish a basis for designing an experimental comparison of the two model types. Differences in steady state effluent concentrations predicted by both models could to a large extent be explained by different model concepts. The steady state system performance was analyzed by evaluating the Monod factor values, and through a sensitivity analysis of the kinetic model parameters. Both methods complement each other. Analysis of the Monod factor values can lead to determination of parameters to be estimated during model calibration. The steady state system response to manipulation of the potential actuators for control was evaluated via a sensitivity analysis. The concept of relative sensitivity was introduced to compare the relative effect of each actuator in both models. The negative relative sensitivities of X S to four of the five control handles analyzed imply an opposite response of both models, which can be important for control structure design. The analysis of the process behavior to different disturbances showed different dynamics of both models. ASM3 simulation results are easier to interpret because the model structure is more transparent, mainly due to the simpler cell decay model principle considered in ASM3. An inverse response was obtained for the return sludge and nitrate recycle flow rate, indicating that multivariable control design is required.

Real-time optimization (RTO) is widely used in industry to improve the steady-state performance of a process using the available measurements, reacting to changing prices and demands scenarios and respecting operating, contractual, and environmental constraints. Traditionally, RTO has used nonlinear continuous formulations to model the process. Mixed-integer formulations have not been used in RTO, because of the need of a fast solution (on the order of seconds or a few minutes), and because many discrete decisions, such as startups or shutdowns, are taken with less frequency in a scheduling layer. This work proposes the use of disjunctions in RTO models, listing a series of examples of discrete decisions (different to startups or shutdowns) that can be addressed by RTO. Two model adaptation approaches (the two-step approach and the modifier adaptation strategy) are revised and modified to make them suitable for RTO with discrete decisions. Some common techniques used in RTO (such as filtering the optimal ...

Abstract An optimization study of membrane-based separation systems for carbon dioxide capture from flue gas of power plants is conducted, considering the possibility of employing up to four stages and using diverse options to create the required driving force. By proposing a superstructure-based model, the number of stages, recycle options, use of feed compression and/or permeate vacuum, driving force distribution along each membrane stage, operating conditions and equipment sizes are simultaneously optimized in order to minimize the total annual cost at high capture ratios and purity targets. Thus, different optimal arrangements are obtained and the total cost is reduced in about 20% compared without employing vacuum. Besides the optimal number of stages diminishes with decreasing purity, but it is independent of the capture ratio. Also, the total cost decreases with the increase of the membrane permeance requiring lower values of operating pressure and membrane areas. Permeance values higher than 2400 GPU lead to lower number of stages and recycles for the same separation target. By contrast, a sensitivity analysis shows that the total cost increases with the increase of the electricity price, capacity factor, and capital recovery factor, which are the more influential parameters in the objective function. Despite new optimal operating and design conditions are obtained when these parameters vary, no modifications in the optimal arrangement are observed.

Abstract This work deals with the development and implementation of mathematical models in the General Algebraic Modeling System (GAMS) environment for optimization purposes, involving extrinsic functions that are executed outside GAMS from dynamic-link libraries (DLL) implemented in the programming language C. Three DLL libraries are developed to calculate thermodynamic properties: the Raoult's law for vapor-liquid equilibrium, the Non-Random Two-Liquid (NRTL) model, and the Peng–Robinson equation of state. A detailed description on how GAMS and DLL libraries interact is presented. Case studies dealing with the optimal design of multi-component distillation columns with increasing complexity levels are discussed. For the proposed case studies, the obtained results show that the usage of the proposed extrinsic functions allows to significantly enhance the model implementation compared to the traditional model implementation approach, and to considerably reduce the model size as well as the computational time required by the optimization algorithms.