The goal of simulating the performance of an expert is to help human workers solve real-world problems by expertise, a specific domain of knowledge (Shiau, 2011). There are diverse problems which need to be solved in the real world. Thus, the use of an expert system (or a similar artificial intelligence framework) becomes prolific in many fields (Liao, 2005). One of the complex problems for the control in which a computational intelligent approach is amenable is a crude oil distillation unit. In a crude distillation process, the first objective is to perform an entire process optimization including high production rate with a required product quality by searching for an optimal operating condition of the operating variables (Frenkel, 2011 and Ouattara et al., 2012). In the previous decade, there was considerable research concerning the optimization of crude distillation processes (Ghashghaee & Karimzadeh, 2011). In Seo, Oh, and Lee (2000), the optimal feed location on both the main column and stabilizer is obtained by solving rigorous “a priori” models and mixed integer nonlinear programming. The sensitivity to small variations in feed composition is studied in Dave, Dabhiya, Satyadev, Ganguly, and Saraf (2003). Julka et al. propose in a two-part paper (Julka et al., 2002a and Julka et al., 2002b) a unified framework for modeling, monitoring and management of supply chain from crude selection and purchase to crude refining. In addition to analytical non-linear models, computational intelligence techniques such as neural networks (Gueddar and Dua, 2012 and Liau et al., 2004) and genetic algorithms (Motlaghi, Jalali, & Ahmadabadi, 2008) are used for the same purpose. Alhajree, Zahedi, Manan, and Zadeh (2011) cite several Artificial Neural Network research studies for the control of processes in petrochemicals and refineries. From cited papers, most of the nonlinear controllers require the feedback of state information for effective control and close monitoring of a process. In practice, however, the complete online information about the present state of the industrial process is rarely available. If the real-world values are not provided to the algorithm on time, the control algorithm becomes formally invalid. In practice, it recovers from the situation, at the price of reduced quality control (i.e., worse product), so such situations should be avoided (Metzger & Polakow, 2011).
The scope of this present study is concerned with a part of the crude oil distillation called the platforming unit. It is made up of two subunits: the catalytic reforming or reaction unit and the distillation unit or train distillation. Most of the cited references are focused on optimizing the production rate of the distillation unit (Iranshahi et al., 2011 and Meidanshahi et al., 2011), but if the focus is the heat recovery, 80% of the energy consumption (67% of the energy invoicing tasks) corresponds to the fuel consumption in the boilers of the previous task (the reaction unit).
At present, research is not only focused in the rise of the production rate but also in making customized products (Frenkel, 2011) and in the improvement of product quality (Rahimpour, Vakili, Pourazadi, Iranshahi, & Paymooni, 2011). In this sense, classic applications of linear control theories on the distillation unit are widely available in the literature (Jabbar & Alatiqi, 1997). Also nonlinear state estimation research (Jana, Samanta, & Ganguly, 2009) and optimal planning strategy research (Kuo & Chang, 2008) are available. The main objective of these papers was to remove impurities in the distillate (i.e., View the MathML sourceC5+ in the debutanizer column) and maintain the minimum possible amount of product (butane) in the bottom residual fuel oil to maximize the yield of the product.
The energy management (de Lima and Schaeffer, 2011 and Kansha et al., 2011) and the energy efficiency (Chiwewe & Hancke, 2011) become important problems. The objective is to perform a complete plant energy process optimization, including an adequate production rate with the required product quality while minimizing operating costs (fuel consumption in boilers) through a data mining approach. Several research endeavors have treated consumption analysis as a knowledge discovery problem using intelligence techniques (Li, Bowers, & Schnier, 2010). Both forms of learning, supervised and unsupervised, have been adopted in these studies (Hippert, Pedreira, & Souza, 2001; Metaxiotis, Kagiannas, Askounis, & Psarras, 2003). In Hippert et al. (2001), the unsupervised learning based on the SOM algorithm for the three tasks, namely classification, filtering and identification of customer load pattern, is proposed. The intelligent control algorithms applied to the control of combustion processes have produced satisfactory results and show a great potential for growth. Previous research has shown that boiler efficiency can be optimized with data-mining approaches (Miyayama et al., 1991 and Ogilvie et al., 1998). In Kusiak and Song (2006), the authors proposed an optimization with clustering-derived centroids. In Song and Kusiak (2007), the authors develop a data mining approach for optimizing the combustion efficiency of an electric-utility boiler subject to industrial operating constraints. The latest cited papers offer interesting researches about single boilers. These studies encourage the authors of the current paper to offer a mining approach to optimize the efficiency of a complete distillation plant, regarding the operating and economical constraints.
Since close monitoring of the process is, in practice, rarely available, only information collected in a historical database and the data mining software tools were used. The expert's performance is hidden in the collected dataset. This valuable knowledge feeds the proposed Decision Support System (DSS) framework. The global plant control model does not need to be reconfigured. The expert's information can simply be extracted.
The questions that emerge are: is it possible to extract expert information from the limited amount of data collected in the historical database, searching in past data optimal cost operating conditions? And, is it possible to improve energy efficiency result by the estimation of new operating condition with a DSS software tool? The feasibility and benefits of the proposed framework are demonstrated with a real case study reported. The proposed framework-based pilot commercial software is also presented.
The paper is organized as follows: in Section 2, the refinery platforming unit process is described. In Section 3, the data mining-based DSS framework is presented. It is divided into four subsections: the nature of the data set, the data preprocessing (cleaning and filtering), the data transformation and discretization and finally, the data reduction and prediction. In Section 4, a solution to increase the plant energy efficiency is proposed. Section 5 illustrates the quality of the framework by a case study considering a real database. In Section 6, a framework-based commercial software is presented. Section 7 outlines future directions and concluding remarks.
