Panos Seferlis

Automatic Synthesis of Alternative Heat-Integrated Water-Using Networks

A novel sequential approach is developed in this work to identify alternative heat-integrated WUNs with four mathematical programming models. The proposed synthesis method can be implemented in two stages. In the first stage, the objective is to identify all alternative optimal configurations featuring the minimum freshwater usage, the minimum energy consumption and the minimum total interconnection number respectively. These structures can be determined by solving two distinct models in sequence. The possible WUN designs are identified exhaustively by adopting the ‘solution pool’ technique provided by Solver CPLEX in GAMS. The HEN designs are then generated in the second stage. From each candidate obtained in the first stage, the stream data can be readily extracted and utilized in the transshipment models for determining the minimum utility usage and the minimum number of heat transfer units. The advantages of this approach are twofold: (1) more than one alternative designs can be found, and (2) all model scales can be kept to manageable levels. A case study is provided in this paper to illustrate the effectiveness of the proposed method.

Computer-Aided Molecular Design: Fundamentals, Methods, and Applications

Introduction 4 Molecular Representation and Property Prediction 5 GC Approach 5 Other Molecular Representations 6 Property Prediction Models 7 Pure component properties 7 Activity coefficient models 7 Cubic equations of state 7 EoS/G models 9 Advanced statistical thermodynamic models—The statistical associating fluid theory approach 10 Computational chemistry models—The conductor-like screening model approach 12 Main Classes of CAMD Problems 13 Classification Overview 13 Single-Objective Computer-Aided Molecular Design 14 Computer-Aided Molecular and Process Design 15 MultiObjective Computer-Aided Molecular Design 15 Computer-Aided Mixture Design 16 Computer-Aided Molecular Design Under Uncertainty 16 Computer-Aided Molecular, Process, and Control Design 16 Computer-Aided Molecular Design for Reactive Systems 18 Computer-Aided Molecular Design Using Quantum Mechanical Models 18 Summary of Classes 18 General Solution Approaches 19 Challenges in the Solution of Molecular Design Problems 19 Classification of Solution Approaches 19 Generate and Test Approaches 20 Deterministic Optimization Algorithms 21 Overview of main algorithms 21 Deterministic optimization approaches in molecular design 23 Stochastic Optimization Algorithms 25 Overview of main algorithmic concepts 25 Genetic algorithms (GA) 25 Simulated annealing 26 Tabu search 27 Ant colony optimization 27 Property Clustering Approaches 27 Merits and Shortcomings of Solution Approaches 28 Methodologies for Solution of CAMPD Problems 30 Generate and Test Approaches 30 Deterministic Optimization Approaches 30 Stochastic Optimization Approaches 33 Continuous Molecular Representations 34 Property Clustering and/or Reverse Formulations 36 Methodologies for Various Problem Classes 36 Multiobjective Optimization 36 Mixture and Blend Design 38 Molecular Design Under Uncertainty 43 Molecular, Process, and Control Design 46 Exploitation of Advanced Computing Infrastructures 47 Methodologies for Reactive Systems or QM-Based Approaches 47 Molecular Design 51

Enhancement of hybrid renewable energy systems control with neural networks applied to weather forecasting: the case of Olvio

In this paper, an intelligent forecasting model, a recurrent neural network (RNN) with nonlinear autoregressive architecture, for daily and hourly solar radiation and wind speed prediction is proposed for the enhancement of the power management strategies (PMSs) of hybrid renewable energy systems (HYRES). The presented model (RNN) is applicable to an autonomous HYRES, where its estimations can be used by a central control unit in order to create in real time the proper PMSs for the efficient subsystems’ utilization and overall process optimization. For this purpose, a flexible network-based design of the HYRES is used and, moreover, applied to a specific system located on Olvio, near Xanthi, Greece, as part of Systems Sunlight S.A. facilities. The simulation results indicated that RNN is capable of assimilating the given information and delivering some satisfactory future estimation achieving regression coefficient from 0.93 up to 0.99 that can be used to safely calculate the available green energy. Moreover, it has some sufficient for the specific problem computational power, as it can deliver the final results in just a few seconds. As a result, the RNN framework, trained with local meteorological data, successfully manages to enhance and optimize the PMS based on the provided solar radiation and wind speed prediction and make the specific HYRES suitable for use as a stand-alone remote energy plant.