فهرست مطالب

  • Volume:6 Issue:2, 2018
  • تاریخ انتشار: 1397/07/09
  • تعداد عناوین: 5
|
  • David Wood *, Abouzar Choubineh Pages 1-14
    Machine-learning algorithms aid predictions for complex systems with multiple influencing variables. However, many neural-network related algorithms behave as black boxes in terms of revealing how the prediction of each data record is performed. This drawback limits their ability to provide detailed insights concerning the workings of the underlying system, or to relate predictions to specific characteristics of the underlying variables. The recently proposed transparent open box (TOB) learning network algorithm successfully addresses these issues by revealing the exact calculation involved in the prediction of each data record.  That algorithm, described in summary, can be applied in a spreadsheet or fully-coded configurations and offers significant benefits to analysis and prediction of many natural gas systems. The algorithm is applied to the prediction of natural gas density using a published dataset of 693 data records involving 14 variables (temperature and pressure plus the molecular fractions of the twelve components: methane, ethane, propane, 2-methylpropane, butane, 2-methylbutane, pentane, octane, toluene, methylcyclopentane, nitrogen and carbon dioxide). The TOB network demonstrates very high prediction accuracy (up to R2 =0.997), achieving comparable accuracy to the predictions reported (R2 =0.995) for an artificial neuralnetwork (ANN) algorithm applied to the same data set. With its high levels of transparency, the TOB learning network offers a new approach to machine learning as applied to many natural gas systems.
    Keywords: Predicting gas density, Learning networks, Multi-component natural gas, Auditable machine learning, Transparent predictions
  • Sakineh Chavoshi *, Mani Safamirzaei, F Pajoum Shariati Pages 15-36
    One of the important, practical and simple methods for hydrate formation conditionis empirical equations, and so far many empirical equations have been presented to predict thetemperature and pressure of hydrate formation. In this study, the methods and empiricalcorrelations have been reviewed and their predictive capabilities have been evaluated with theuse of more than 2000 experimental data collected from literature. These data have beenseparated in three groups: (1) simple natural gas components included methane, ethane,propane, isobutane, carbon dioxide, nitrogen, and hydrogen sulfide (2) binary gas mixturesand (3) gas mixture similar to natural gas. In this paper, after expressing the restrictions ofsome empirical correlations have been proposed by scientists before and proposed empiricalcorrelation in the present study, the results of evaluating have been presented in severaltables and curves. The proposed empirical correlation in the present study has shown reliableperformance for both simple natural gas components and mixtures. Despite the existencethree adjustable parameters, the accuracy of this equation shows the ranking 1 to 3 compareto the rest of the equations.
    Keywords: Empirical Correlations, Gas Hydrate Formation Condition, Natural Gas Mixtures, Hydrate Formation Predicting, Hydrate Formation Temperature
  • Mosayeb Paashang, Seyyed Mohammad Sharifi *, Gholamreza Salehi Pages 37-48
    One of the industries with high potential for energy saving is the petrochemical industry. Ethylene and propylene production plants (olefin plants) – as a part of the petrochemical industry – are very energy intensive. So, any try to improve their energy consumption efficiency could lead to a high amount of energy saving. Iran’s petrochemical industry uses old technologies and components and due to sanctions, it couldn’t be improved. The main idea of this paper is to improve the energy consumption of one of the biggest petrochemical plants in Iran. So, Marun olefin plant in Iran has been simulated as a case study and its different parts have been analyzed from exergy point of view, which shows the most energy intensive components so that we can focus on for improving the plant’s energy consumption. The plant has been divided into three sections and simulated using Aspen HYSYS process simulation software. Then, it has been analyzed using exergy analysis. Results show that the hydrogenation and separation section consisting of many different components has the highest exergy destruction rate and the highest potential for energy saving. Compression section and refrigeration system having compressors are the other parts highly destroying exergy respectively. The causes of exergy destruction for each component has been analyzed and recommendations have been proposed as well
    Keywords: Olefin, Ethylene Plant, Exergy Analysis, Exergy Destruction, Exergetic Efficiency
  • Parisa Naeiji, Farshad Varaminian * Pages 49-60
    In this study, a differential scanning calorimetry (DSC) has been used to characterize tetrahydrofuran (THF) hydrate formation with and without the presence of sodium and chloride salts. The thermal properties including the heat of formation, phase change temperature and specific heat capacity of THF hydrate formation have been determined. When salts are present, THF hydrate is inhibited, so that the phase change temperature shifts to a lower temperature, a broadening of the DSC peak can be seen and the height of the peak and the heat of formation decreases. Because the heat capacity depends on the heat of formation, it reduces when salt is present. The results show that the sodium salt solutions have exhibited the best performance on delaying nucleation of hydrate in order of Na2SO3 > NaF > NaCl, while chloride salt solutions, especially CaCl2, have demonstrated that can reduce the heat of formation to around 65% compared to that of without salt.
    Keywords: Tetrahydrofuran Hydrate, Differential Scanning Calorimetry, Formation Heat, Phase Change Temperature, Specific Heat Capacity
  • Alireza Behroozsarand *, Kamran Ghasemzadeh, Hossein Soltanalizadeh Pages 61-74

    In this paper, an industrial ethylbenzene production unit and two new reactor arrangements have been simulated and the results were compared. In most petrochemical plants, ethylene as ethylbenzene production feed is produced at steam thermal cracking units. Almost ethylene is produced excess in these plants because of changing feed rate, feed type and or shut down of any other user unit of ethylene. So finding a new application for consuming excess ethylene in these plants without designing new units is important. In this study, two new reactor arrangements have been proposed. In the first scenario, the 3rd bed of transalkylator plays alkylator role instead of transalkylator and in the second scenario, all three beds were used as a parallel reactor with current alkylator reactors. The results show that ethylbenzene productivity rises 22% and 20% in scenarios 1 and 2, respectively, compared to a commercial industrial unit. Finally, one of the effective parameters in ethylbenzene unit is the ratio of ethylene to benzene in the feed of all beds of reactors, so it was selected as decision variable and ethylbenzene productivity as the objective function of an optimization problem. All cases were optimized and results show 43% ethylbenzene productivity improvement in an optimized version of scenario2 in compared other scenarios.

    Keywords: Ethylbenzene, Optimization, Alkylation, New arrangement, Productivity