چارچوب و سیستم بهینه سازی شبیه سازی مبتنی بر دانش برای عملیات فرایند پایدار

Design and operation of chemical plants involves a combination of synthesis, analysis and evaluation of alternatives. Such activities have traditionally been driven by economic factors first, followed by engineering, safety and environmental considerations. Recently, chemical companies have embraced the concept of sustainable development, entailing renewable feed materials and energy, non-toxic and biodegradable products, and waste minimization or even elimination at source. In this paper, we introduce a knowledge-based simulation-optimization framework for generating sustainable alternatives to chemical processes. The framework has been developed by combining different process systems engineering methodologies – the knowledge-based approach for identifying the root cause of waste generation, the hierarchical design method for generating alternative designs, sustainability metrics, and multi-objective optimization – into one coherent simulation-optimization framework. This is implemented as a decision-support system using Gensym's G2 and the HYSYS process simulator. We illustrate the framework and system using the HDA and biodiesel production case studies.

The notion of sustainable development – “development that meets the needs of the present without compromising the ability of future generations to meet their own needs” (WCED, 1987) – has prompted numerous actions from governments, businesses, institutions, and industries worldwide to balance economic activities with environmental and social responsibilities. In Finland, for example, sustainable development has become a central element in driving the government policies towards improving the life of its citizens (Finland's Environmental Administration, 2006). In the U.S., various initiatives have been launched by technical institutions including the Institute for Sustainability (American Institute of Chemical Engineers) and the Green Chemistry Institute (American Chemical Society) to promote sustainable products and processes (Beloff & Lines, 2005). The expanding commitment of the industrial sector is also evident from their annual sustainability targets and achievements (BP, 2005 and Shell, 2005). All these show the impact of sustainable development concept on various spheres of human activities. Given their role as a large-scale provider of material goods within society, the chemical industries consume large amounts of non-renewable resources and concomitantly emit wastes. Representing 4% of the world economy, the chemical processing plants, with a global turnover of €1 841 billions and 10 million employees, is currently responsible for 7% of global energy use (Lines, 2005), and 4% of the total CO2 emissions to the atmosphere (Jenck, Agterberg, & Droescher, 2004). They therefore, have an important role in contributing toward sustainable development. Specifically, to enhance their long-term sustainability, various environmental considerations including reducing raw material and energy usage, switching to renewable feedstock, and waste reuse and recycling needs to be implemented. Certainly, such measures would require changes to existing processes – ranging from simple modifications of the design and operation to more intrusive options such as material substitution and technology upgradation. Several techniques can be used to identify opportunities for reducing pollutant generation as well as material and energy consumption within a process plant, including industrial ecology, life-cycle assessment (LCA), green chemistry, and waste minimization. These four techniques are not mutually exclusive but each seeks to improve the sustainability of a plant from a different perspective. At a geographical cluster level, industrial ecology is a method to improve the environmental impact of a plant through waste exchange, recycle, and reuse with other plants in the vicinity (Ehrenfeld & Gertler, 1997). One example of the successful implementation of this technique is at the Kalundborg industrial park in Denmark (Chertow, 2000), where an oil refinery, power station, gypsum board facility, pharmaceutical plant, and the city itself, share water, steam, and electricity resources, and also exchange a variety of wastes. The outcome is a 25% reduction of the fresh water usage, 2.9 million tons of material recycling, and energy for heating 5 000 homes. LCA is a tool for elucidating the environmental burdens over the entire life-cycle of the product, starting from raw material extraction to production process, point of use, and final disposal (SETAC, 1994). Although LCA traditionally focuses on products and their impacts on the environment, it has also been applied as a decision-making tool during process design (Kniel, Delmarco, & Petrie, 1996). While industrial ecology and LCA focus outwards from the process and plant, green chemistry and waste minimization look inwards. The production process can be made inherently benign through green chemistry, which involves designing new processes or products (such as catalysts) that eliminate or reduce the use and generation of hazardous substances (Anastas & Warner, 1998). Given its nature, this is mostly applicable in the initial design stages where changes to the process chemistry are still viable. On the other hand, waste minimization is a manufacturing-centric activity which avoids, eliminates or reduces waste at its source, or allows reuse or recycling of the waste within a plant (Crittenden & Kolaczkowski, 1995). It is thus suited for initial process design as well as the retrofit situation, where different modifications can be proposed to the base case design and operation in order to improve the environmental performance. This paper presents a knowledge-based simulation-optimization framework for generating sustainable design and operations alternatives for chemical process plants. The proposed framework, which has been implemented as a fully automated decision-support system, combines different process systems engineering (PSE) methodologies – knowledge representation, waste source diagnosis, knowledge-based design retrofitting, quantitative assessment of the alternatives and multi-objective optimization – for sustainable process operations problem. The paper is organized as follows: in the next section, we review the literature on waste minimization. In Section 3, the knowledge-based simulation-optimization based framework is proposed. The integration of the heuristic-based waste diagnosis with process simulation and mathematical optimization is described in Section 4. The development of a decision-support system that automates the various elements of the framework is also outlined. In Section 5, we illustrate the framework on a biodiesel production process

The need for sustainable development has challenged the chemical process industries to seek new approaches to tackle the waste problem. This includes exploitation of popular, commercial tools such as process simulators to evaluate process retrofit options. While process simulators are useful, their application to the waste minimization is not straightforward as considerable knowledge, skill and expertise are required in the part of the user to identify the key units and variables that control the overall waste feature of the plant. In this paper, we present a knowledge-based simulation-optimization framework comprising of an expert system, process simulator, and multi-objective optimization for the purpose of waste minimization analysis. The framework has been developed by capitalizing on the expert system's ability to automatically perform knowledge-based waste source diagnosis and qualitative design solution identification. The functional model of the process reveals the relevant design and operation variables that should be manipulated to implement these alternatives. The optimal manipulation is identified using simulated annealing algorithm which uses the process simulator to evaluate the environmental and economic objectives. The entire framework has been automated so that it is transparent to the user. ENVOPExpert can be used by any process engineer even without specialized waste minimization or optimization knowledge. The framework and tool has been successfully tested using two case studies: HDA process and biodiesel production. While the overall results from ENVOPExpert are very encouraging, the design alternative analysis and evaluation methods implemented in ENVOPExpert can be further improved. Due to its qualitative nature, the P-graph analysis cannot rank the waste sources or the design alternatives based on the potential for improvement or identify trade-offs between conflicting effects. An alternative analysis technique that is based on indicator metrics, as implemented in SustainPro (Carvalho et al., 2008), identifies the trade-offs between various effects and ranks various retrofit options. An integration of ENVOPExpert and SustainPro would allow quantification of the cost-benefit of the design alternatives, lead to more accurate analysis, and more focused optimization of design variables. This will be a focus of our further research. In the future, we will also extend the framework to batch process operation (Halim & Srinivasan, 2006). In addition, we intend to integrate utilities (such as water and energy) minimization into the framework. Incorporating the cost of CO2 emission – at 85 $/t (Hardisty, 2007) – into the economic analysis will also be considered in the future. This will significantly enhance the objective of ENVOPExpert as a decision-support system for sustainability studies at various stages of the plant life-cycle.