Research Outputs
Permanent URI for this communityhttps://hdl.handle.net/20.500.14288/2
Browse
9 results
Search Results
Publication Metadata only A dynamic non-isothermal model for a hydrocracking reactor: model development by the method of continuous lumping and application to an industrial unit(Elsevier Sci Ltd, 2012) Çakal, Berna; Gökçe, Dila; Kuzu, Emre; N/A; Department of Chemical and Biological Engineering; Şıldır, Hasan; Arkun, Yaman; Phd Student; Faculty Member; Department of Chemical and Biological Engineering; Graduate School of Sciences and Engineering; College of Engineering; 242076; 108526Hydrocracking is an important refinery process which is carried out in catalytic reactors to convert heavy petroleum fractions into valuable products. Because of the large number of species and complex reactions involved, modeling of hydrocracking is a challenging task. In this paper a dynamic, non-isothermal reactor model has been constructed using the method of continuous lumping which treats the complex reactive mixture as a continuum. In doing so concentrations are characterized in terms of reactivity which is a monotonic function of the true boiling point of the mixture. The material and energy balances are developed in the form of integro-differential equations. The significant modeling parameters are identified and estimated using data from an industrial reactor. Steady-state and dynamic predictions of the model outputs such as reactor temperature, product yields and hydrogen consumption are shown to be in good agreement with plant data.Publication Metadata only Design and adaptive sliding-mode control of hybrid magnetic bearings(Institute of Electrical and Electronics Engineers (IEEE), 2018) N/A; N/A; Department of Mechanical Engineering; Zad, Haris Sheh; Khan, Talha Irfan; Lazoğlu, İsmail; PhD Student; PhD Student; Faculty Member; Department of Mechanical Engineering; Graduate School of Sciences and Engineering; Graduate School of Sciences and Engineering; College of Engineering; N/A; N/A; 179391In this paper, a hybrid magnetic bearing (HMB) prototype system is designed and analyzed. Two compact bearings are used to suspend the rotor in five degrees of freedom. Electromagnets are used for axial suspension of the rotor, while permanent magnets are used for the passive radial stability. A brushless DC motor is designed in order to rotate the shaft around its axis. The 3-D finite-element model of the HMB system is established and distribution of magnetic fields in the air gaps and magnetic forces on the rotor under various control currents and displacements is calculated. A nonlinear adaptive sliding-mode controller is designed for the position control of the rotor in axial direction. Since the control characteristics of the active magnetic bearing system are highly nonlinear and time varying with external interference, a radial basis function compensator is designed first, and then, a sliding-mode control law is used to generate the control input. The stability analysis for the designed controller is given based on the Lyapunov theorem. Experimental setup is built to guide the design process. The performance of the HMB system based on the designed control algorithm is evaluated under different operating conditions.Publication Metadata only Detection and mitigation of targeted data poisoning attacks in federated learning(Institute of Electrical and Electronics Engineers Inc., 2022) Department of Computer Engineering; Department of Computer Engineering; Gürsoy, Mehmet Emre; Erbil, Pınar; Faculty Member; Student; Department of Computer Engineering; College of Engineering; College of Engineering; 330368; N/AFederated learning (FL) has emerged as a promising paradigm for distributed training of machine learning models. In FL, several participants train a global model collaboratively by only sharing model parameter updates while keeping their training data local. However, FL was recently shown to be vulnerable to data poisoning attacks, in which malicious participants send parameter updates derived from poisoned training data. In this paper, we focus on defending against targeted data poisoning attacks, where the attacker's goal is to make the model misbehave for a small subset of classes while the rest of the model is relatively unaffected. To defend against such attacks, we first propose a method called MAPPS for separating malicious updates from benign ones. Using MAPPS, we propose three methods for attack detection: MAPPS + X-Means, MAPPS + VAT, and their Ensemble. Then, we propose an attack mitigation approach in which a "clean"model (i.e., a model that is not negatively impacted by an attack) can be trained despite the existence of a poisoning attempt. We empirically evaluate all of our methods using popular image classification datasets. Results show that we can achieve > 95% true positive rates while incurring only < 2% false positive rate. Furthermore, the clean models that are trained using our proposed methods have accuracy comparable to models trained in an attack-free scenario.Publication Metadata only Experimental analysis of boring process on automotive engine cylinders(Springer, 2010) Özkeser, Salih O.; N/A; Department of Mechanical Engineering; Şenbabaoğlu, Fatih; Lazoğlu, İsmail; Master Student; Faculty Member; Department of Mechanical Engineering; Graduate School of Sciences and Engineering; College of Engineering; N/A; 179391In this article, mechanics of boring process on cast iron automotive engine cylinders is explored experimentally. In order to shorten the boring cycle time and to improve quality of the cylinder holes, effects of various cutting conditions as spindle speed, feedrate, inserts, and coatings are investigated. Real-time cutting forces are measured with dynamometer during the process. Surface roughness on the engine cylinders, flank, and crater tool wears are measured and compared in various cutting conditions. It is concluded that by selecting proper cutting conditions, cutting forces can be controlled below a threshold value, cycle time can be shortened, tool life and part quality can be increased; therefore, the cost of automotive engine boring process can be reduced significantly.Publication Metadata only Generating robot/agent backchannels during a storytelling experiment(Institute of Electrical and Electronics Engineers (IEEE), 2009) Al Moubayed, S.; Baklouti, M.; Chetouani, M.; Dutoit, T.; Mahdhaoui, A.; Martin, J. -C.; Ondas, S.; Pelachaud, C.; Urbain, J.; Department of Mechanical Engineering; Yılmaz, Mustafa Akın; Tekalp, Ahmet Murat; PhD Student; Department of Mechanical Engineering; Graduate School of Sciences and Engineering; N/AThis work presents the development of a real-time framework for the research of Multimodal Feedback of Robots/Talking Agents in the context of Human Robot Interaction (HRI) and Human Computer Interaction (HCI). For evaluating the framework, a Multimodal corpus is built (ENTERFACE_STEAD), and a study on the important multimodal features was done for building an active Robot/Agent listener of a storytelling experience with Humans. The experiments show that even when building the same reactive behavior models for Robot and Talking Agents, the interpretation and the realization of the behavior communicated is different due to the different communicative channels Robots/Agents offer be it physical but less-human-like in Robots, and virtual but more expressive and human-like in Talking agents.Publication Metadata only Hybrid systems: modeling, simulation and optimization(Elsevier Sci Ltd, 2009) Karasözen, Bülent; Biegler, Lorenz T.; McAvoy, Thomas J.; Department of Industrial Engineering; Türkay, Metin; Faculty Member; Department of Industrial Engineering; College of Engineering; 24956N/APublication Metadata only Physical activity recognition using deep transfer learning with convolutional neural networks(Institute of Electrical and Electronics Engineers Inc., 2022) Department of Electrical and Electronics Engineering; Department of Computer Engineering; N/A; N/A; Gürsoy, Beren Semiz; Gürsoy, Mehmet Emre; Ataseven, Berke; Madani, Alireza; Faculty Member; Faculty Member; Master Student; Master Student; Department of Electrical and Electronics Engineering; Department of Computer Engineering; College of Engineering; College of Engineering; N/A; Graduate School of Sciences and Engineering; 332403; 330368; N/A; N/ACurrent wearable devices are capable of monitoring various health indicators as well as fitness and/or physical activity types. However, even on the latest models of many wearable devices, users need to manually enter the type of work-out or physical activity they are performing. In order to automate real-time physical activity recognition, in this study, we develop a deep transfer learning-based physical activity recognition framework using acceleration data acquired through inertial measurement units (IMUs). Towards this goal, we modify a pre-trained version of the GoogLeNet convolutional neural network and fine-tune it with data from IMUs. To make IMU data compatible with GoogLeNet, we propose three novel data transform approaches based on continuous wavelet transform: Horizontal Concatenation (HC), Acceleration-Magnitude (AM), and Pixelwise Axes-Averaging (PA). We evaluate the performance of our approaches using the real-world PAMAP2 dataset. The three approaches result in 0.93, 0.95 and 0.98 validation accuracy and 0.75, 0.85 and 0.91 test accuracy, respectively. The PA approach yields the highest weighted F1 score (0.91) and activity-specific true positive ratios. Overall, our methods and results show that accurate real-time physical activity recognition can be achieved using transfer learning and convolutional neural networks.Publication Metadata only Plant-wide hierarchical optimization and control of an industrial hydrocracking process(Elsevier Sci Ltd, 2013) Çakal, Berna; Gökçe, Dila; Kuzu, Emre; N/A; Department of Chemical and Biological Engineering; Şıldır, Hasan; Arkun, Yaman; PhD Student; Faculty Member; Department of Chemical and Biological Engineering; Graduate School of Sciences and Engineering; College of Engineering; 242076; 108526Hydrocracking is a crucial refinery process in which heavy hydrocarbons are converted to more valuable, low-molecular weight products. Hydrocracking plants operate with large throughputs and varying feedstocks. In addition the product specifications change due to varying economic and market conditions. In such a dynamic operating environment, the potential gains of real-time optimization (RTO) and control are quite high. At the same time, real-time optimization of hydrocracking plants is a challenging task. A complex network of reactions, which are difficult to characterize, takes place in the hydrocracker. The reactor effluent affects the operation of the fractionator downstream and the properties of the final products. In this paper, a lumped first-principles reactor model and an empirical fractionation model are used to predict the product distribution and properties on-line. Both models have been built and validated using industrial data. A cascaded model predictive control (MPC) structure is developed in order to operate both the reactor and fractionation column at maximum profit. In this cascade structure, reactor and fractionation units are controlled by local decentralized MPC controllers whose set-points are manipulated by a supervisory MPC controller. The coordinating action of the supervisory MPC controller accomplishes the transition between different optimum operating conditions and helps to reject disturbances without violating any constraints. Simulations illustrate the applicability of the proposed method on the industrial process.Publication Metadata only Robotic additive turning with a novel cylindrical slicing method(Springer London Ltd, 2022) N/A; N/A; Department of Mechanical Engineering; Yiğit, İsmail Enes; Khan, Shaheryar Atta; Lazoğlu, İsmail; Phd Student; PhD Student; Faculty Member; Department of Mechanical Engineering; Graduate School of Sciences and Engineering; Graduate School of Sciences and Engineering; College of Engineering; N/A; N/A; 179391The turning process used from ancient times to today's modern turning centers is based on material removal. This article presents a new work to integrate additive manufacturing into the turning process and generate complex free-form additive turning part geometries. The conventional slicing method used in AM is the planar slicing method. In the planar slicing method, the computer-aided design (CAD) model is sliced using planes, and as a result, two-dimensional toolpaths are formed. A new slicing method is required to achieve additive turning parts. This work proposes a generalized, cylindrical slicing method that generates nonplanar toolpaths wrapped around a cylinder. The model is sliced by cylindrical layers, with increasing radii at each layer. As a result, three-dimensional toolpaths that are suitable for additive turning are generated. In conventional AM, lower tensile strength is observed in the build orientation of the part where the layers bind. Additive turning increases the low tensile strength observed in conventional AM. Additionally, it reduces and, at times, even eliminates the support structures required for certain CAD models. The cylindrical slicing results are verified by additively turning different CAD models using a six-axis robotic serial manipulator fitted with a fused filament fabrication end effector and an external turning axis. Tensile tests are conducted on conventional AM and additive turning models to verify the improvement in tensile strength.