Sharif Digital Repository

Semantic segmentation is the tash of labeling each pixel of an input image. It is one of the main problems in computer vision and plays an important role in scene understanding. State of the art methods of solving it are based on Convolutional Neural Networhs (CNNs). While many real world tasks like autodriving cars and robot navigation require fast and lightweight models, CNNs inherently tend to give beter accuracy when they are deeper and bigger, and this has raised interest in designing compact networks. Knowledge distillation is one of the popular methods of training compact networhs and helps to transfer a big and powerful network’s knowledge to a small and compact one. In this research... 

In this study steady state and dynamic simulation of an NFM extractive distillation unit for seperation of aromatics and non-aromatic components has been carried out. Column output composition obtained with experimental binary coefficients is compared with those obtainded by UNIFAC equation. The results indicate that using UNIFAC for all binrary coefficients results in an unacceptable error in concentration of benzene. If the experimental binary coefficents of benzene with NFM are used and the rest are calculated by UNIFAC equation, the simulation results will be very close to those obtained via experimental binary coefficients. Also it is shown that considering a constant efficiency for all... 

Nowadays, neural networks are used as the main method in most machine learning applications. But research has shown that these models are vulnerable to adversarial attacks imperceptible changes to the input of neural networks that cause the net- work to be deceived and predict incorrectly. The importance of this issue in sensitive and security applications of neural networks, such as self-driving cars and medical diagnosis systems, becomes much higher. In recent years, many researches have been done in the field of making neural net- works robust against this threat, but in most of them, higher robustness has been provided on the basis of larger and more complex models. Few researches have... 

Simulation and controlling nonlinear process is one of the challenging problems in academic and industrial research. Saving data for keeping history of process or arising number of calculation for optimization in control design, is forcing to decrease number of equation or order of system. One of the methods to decrease the order of distillation process is using the travelling wave theory. This solution shows god accuracy for simulation the behavior of distillation column. On the other hand finite difference method as a solver for distillation equation takes some time to converge to steady state condition and the numbers of calculation in this method are high. Finite difference method is... 

In a polyvinyl alcohol (PVA) plant, equal moles of methyl acetate are produced as byproduct for every mole of PVA. Since there is a little demand for impure methyl acetate, it is converted to methanol and acetic acid as raw materials of PVA plant via hydrolysis reaction. Duo to the small equilibrium constant of hydrolysis reaction of methyl acetate, the conventional process (including reaction and separation sections) suffers some problems such as large recycle flow and high energy consumption. Therefore, reactive distillation is introduced as an alternative for these processes. Since the reactive distillation integrates both reaction and separation in a single vessel, it has more complex... 

Natural boron consists of two stable isotopes, mass 10 (19.3%) and mass 11 (80.7%). 10B isotope has a large thermal neutron capture cross-section and it therefore, used widely in such aspects as modern industry, military equipment, medical science and so on. specially the demand of atomic energy industry for the boron isotope is increased severely. So the separation of 10B from natural boron has received very wide attention at present. There are many kinds of methods on the boron isotopes separation, such as chemical exchange reaction and distillation in boron trifluoride method, ion exchange in boric acid solution method, counter current recycle membrane cascades in boron trifluoride... 

Given the environmental benefits of natural gas (NG) in comparison to other fossil fuels and its potential contribution to the shift towards renewable energy, there is an anticipated rise in the demand for NG as an energy source in the coming years. This surge in demand has sparked a growing interest in the exploration and development of unconventional gas fields with low hydrocarbon content, which were previously considered economically unviable. Among these fields, those rich in nitrogen have garnered significant attention. However, the presence of high nitrogen content in NG poses notable challenges, including a decrease in calorific value and increased costs associated with... 

Process optimization is a technique by which the plant operating condition is obtained such that a specific objective is achieved while the other process requirements are still met. These requirements can include product quality specifications, physical equipment limitations, and the constraints corresponding to the environmental, health, and safety issues. Frequently, this objective is an economic one. In the process optimization, the use of improper optimization techniques may lead to impractical results. One of these strategies is the plant wide optimization. In order to show the effect of plant wide optimization, one should use a complicated enough process as a bench mark of the... 

Selecting an appropriate control structure is an important subject in the design of control systems. Distillation column is one of the most popular and widely used unit operations in industry which complete understanding of its operability and controllability is one of the most important issues in the process control engineering. The present work studies the method of Output Controllability Index for nonlinear systems and compares it with Relative Gain Array and Condition Number. By using this index, the extent of intrinsic controllability of a process can be identified quantitatively in the absence of any controller. Also, the controllability of the process can be stated numerically or... 

Fossil fuel usage caused global warming and other environmental problems. Recently, biofuels are considered as an important alternative. However, the efficient usage of water and energy in their production may accelerate their diffusion. ETBE is a kind of oxidizing fuel additive that can be biologically produced from the ethanol reaction. High water and energy consumption in the production and separation process is still a challenge. In this work, the ETBE production from lignocellulosic feed are considered. Then, the effect of different parameters are investigated by performing sensitivity analysis. In the next step, the optimal pathway of water and energy use in the production of ETBE are... 

Sarcheshme Copper refinery is using ion exchange columns to supply deionized water to the refinery. Every day, 240 cubic meters of waste water from ion exchange columns process are released in nature. The main purpose of this research is to use membrane distillation to purify this wastewater, which reduces the fresh water requirement and at the same time prevents environmental pollution. Membrane distillation technology is capable to purify the high salinity wastewater that other conventional membrane technologies are not able to purify. In this study, the PTFE membrane was used due to its high permeability and low surface energy with pore size of 0.22 μm and membrane thickness of 0.2 to... 

The detection of unusual events, which is called abnormality, is very important in various fields such as: industry, medicine, art, and agriculture, and has applications such as food quality detection, inconsistency detection in work environments, disease detection in medical images, and artefact detection. It has the art of counterfeiting and detection of unhealthy agricultural products, and by detecting the abnormality, the damages caused by the abnormality can be reduced. There are many challenges when detecting anomalies in images, which can be mentioned: the input image is not rich enough to learn a suitable representation, there are not enough samples to learn the model, the high... 

The current work studies the economic investigation of reverse osmosis (RO) desalination unit and multi-effect distillation (MED) considering the average global and also Iran circumstances. Technical and operational parameters are taken into account for economic modeling of both methods to choose the proper and economic method for water desalination based on different operational conditions. The cost of the desalination plants is investigated in two different parts of capital cost and operating and maintenance cost. Capital cost is divided into direct and indirect cost subsections. Based on these calcifications, the effect of economic parameters on change of capital cost and operating and... 

The labeling process is one of the most important tasks in the field of computer vision. The dense labeling problem is the main step towards 2D and 3D scene understanding. The main goal of dense labeling is to label all pixels of images that are known as a semantic segmentation of images in the related literature. Although the state-of-the-art results are mainly achieved by deep learning methods, traditional methods had also been at the center of attention for some years. In the last decades, convolutional neural networks have changed the landscape of visual recognition tasks such as labeling and semantic segmentation. The most important issues in deep learning models are the hardware and... 

In a single chemical plant, the operation of each unit operation is not independent of each other. In the other hand, the main operations are performed in reactors following by several separation steps including distillation and purification units. The total energy requirements are sullied by the heat exchanger networks. The performance of these systems is improved by integration and intensification of the reactor and distillation column in a single unit. The main objective of the present work is to provide an effective design approach to intensify the reactor and distillation column. In a reactive distillation column, the reaction and separation is done in a single unit. Increasing... 

Cyclic distillation can bring new life in old distillation columns, by using a periodic operation mode that leads to significant benefits, such as increased column throughput, lower energy requirements and much higher separation performance. The design method for this type of column are reviewed. Among two design approaches, the Lagrangian one is selected, where equations are formulated as delay differential equations, moving from bottom to top of the column to determine feed tray location and total tray numbers. No rigorous design method is selected in order to calculate the column parameters.In this thesis, by defining an economic index that encompasses capital investment, operating costs... 

Helium is a valuable substance that is widely used in industry and medicine because of its unique properties. Due to the increasing demand for helium in the global market, several helium extraction units have been launched in the world in recent years; Nevertheless, there is still a significant gap between projected global helium production capacity and demand. Today, natural gas is known as the main source of helium on the earth. The method of industrial extraction of helium from natural gas is cryogenic. Due to the low temperature of this process, relatively high power consumption is used to extract helium from natural gas. Therefore, the optimization of this process is of great... 

Natural boron includes two stable isotopes 10B and 11B with 19.3 and 80.7 weight percent respectivley.10B isotope has a high thermal neutron absorption cross section. So isotope 10B material is suitable for making atomic reactor control rods. Other and even more important applications of 10B are in disclosing neutrons in order to measure neutron reactors fluxes and also are in nuclear physics laboratories and medical radiation. This isotope also is used in the pharmaceutical and the military industries. There are several methods for separating boron isotopes. Among the distillation methods, thermal diffusion of BF3, distillation of BF3, distillation of methyl borate, boric acid distillation... 

The existing continual learning methods are mainly focused on fully-supervised scenarios and are still not able to take advantage of unlabeled data available in the environment. Some recent works tried to investigate semi-supervised continual learning (SSCL) settings in which the unlabeled data are available, but it is only from the same distribution as the labeled data. This assumption is still not general enough for real-world applications and restricts the utilization of unsupervised data. In this work, we introduce Open-Set Semi-Supervised Continual Learning (OSSCL), a more realistic semi-supervised continual learning setting in which out-of-distribution (OoD) unlabeled samples in the... 

Azeotropic separation has multiple difficulties. In order to remove the azeotropes, different solutions are recommended such as the use of entrainers, extractive distillation, and pervaporation. Based on initial condition of the mixture, a decision tree is proposed. As a sample, the Ethanol – Ethyl Acetate system is selected and a new solvent – entrainer is proposed with addition of different salts. The suitable concentrations of salt, entrainers or solvent are tested experimentally to draw the relevant RCM.
A suitable set up is designed and assembled, and then samples are taken to analyze in order to provide T-XY diagram. The pressure also is taken into account to correct its effect....