Please login first
Kaamran Raahemifar      
Timeline See timeline
Kaamran Raahemifar published an article in November 2018.
Top co-authors See all
Sridhar Krishnan

374 shared publications

Electrical and Computer Engineering, Ryerson University Faculty of Engineering and Architectural Science, Toronto, Ontario, CANADA

Ahmad Akbari

204 shared publications

Institute of Nanoscience and Nano Technology, University of Kashan, Kashan, Iran

A. Akbari

105 shared publications

Swansea University Medical School, Swansea, United Kingdom

Bijan Raahemi

59 shared publications

University of Ottawa, Canada

Mahmood Fathy

49 shared publications

Computer Engineering Department; Iran University of Science and Technology; Tehran Iran

Publication Record
Distribution of Articles published per year 
(2012 - 2018)
Total number of journals
published in
Publications See all
Article 0 Reads 0 Citations Design, modeling and economic performance of a vertical axis wind turbine Sahishnu R. Shah, Rakesh Kumar, Kaamran Raahemifar, Alan S. ... Published: 01 November 2018
Energy Reports, doi: 10.1016/j.egyr.2018.09.007
DOI See at publisher website ABS Show/hide abstract
Vertical Axis Wind Turbine (VAWT) is relatively simple to implement in urban areas on ground or/and building-roofs, the development of appropriate design of VAWT will open new opportunities for the large-scale acceptance of these machines. The primary objective of this research was to design and modeling of a small-scale VAWT, which can be used to meet the power for low demand applications. Two new shapes of Savonius rotor blades were examined in terms of their rotational performances against the conventional straight and the curved blades. MATLAB simulation was utilized to develop a mathematical model, which comprised of wind power coefficient, tip speed ratio, mechanical and electrical subcomponents. The measured results of developed turbine were used for the validation of the model. The aims were to analyze the turbine blade shapes, develop a mathematical algorithm, and to establish the techno-economic performance of the new curved shape design. It was modeled that the proposed turbine is capable of producing an annual energy output of 7838 kWh and the annual electricity cost/saving in Ontario turned out to be $846.51 (the price of electricity was taken $0.108/kWh).
Article 0 Reads 0 Citations Co-benefit analysis of incentives for energy generation and storage systems; a multi-stakeholder perspective Ehsan Haghi, Michael Fowler, Kaamran Raahemifar Published: 01 September 2018
International Journal of Hydrogen Energy, doi: 10.1016/j.ijhydene.2018.08.150
DOI See at publisher website
Article 0 Reads 1 Citation From a 3D Passive Biped Walker to a 3D Passivity-Based Controlled Robot Borhan Beigzadeh, Mohammad Reza Sabaapour, Mohammad Reza Hai... Published: 01 August 2018
International Journal of Humanoid Robotics, doi: 10.1142/s0219843618500093
DOI See at publisher website
Article 6 Reads 0 Citations An Automatic Image Processing System for Glaucoma Screening Ahmed AlMazroa, Sami Alodhayb, Kaamran Raahemifar, Vasudevan... Published: 01 January 2017
International Journal of Biomedical Imaging, doi: 10.1155/2017/4826385
DOI See at publisher website ABS Show/hide abstract
Horizontal and vertical cup to disc ratios are the most crucial parameters used clinically to detect glaucoma or monitor its progress and are manually evaluated from retinal fundus images of the optic nerve head. Due to the rarity of the glaucoma experts as well as the increasing in glaucoma’s population, an automatically calculated horizontal and vertical cup to disc ratios (HCDR and VCDR, resp.) can be useful for glaucoma screening. We report on two algorithms to calculate the HCDR and VCDR. In the algorithms, level set and inpainting techniques were developed for segmenting the disc, while thresholding using Type-II fuzzy approach was developed for segmenting the cup. The results from the algorithms were verified using the manual markings of images from a dataset of glaucomatous images (retinal fundus images for glaucoma analysis (RIGA dataset)) by six ophthalmologists. The algorithm’s accuracy for HCDR and VCDR combined was 74.2%. Only the accuracy of manual markings by one ophthalmologist was higher than the algorithm’s accuracy. The algorithm’s best agreement was with markings by ophthalmologist number 1 in 230 images (41.8%) of the total tested images.1. IntroductionAs the world’s population has drastically increased, the number of people suffering from glaucoma, or those suspected to have glaucoma, has increased too. Therefore, there is an even greater need for proper diagnosis and effective control of glaucoma. Accurate diagnosis of glaucoma requires three different sets of examinations: (1) evaluation of the intraocular pressure (IOP), (2) evaluation of the visual field, and (3) evaluation of the optic nerve head [1]. Since both elevated-tension glaucoma and normal-tension glaucoma may or may not increase the IOP, the IOP by itself is not a sufficient screening or diagnosis method [2]. On the other hand, visual field examination requires special equipment which is usually available only in tertiary care hospitals equipped with a fundus camera, parametric instrumentation, and possibly an optical coherence tomography [2]. The optic nerve head examination (cup to disc ratio) is a valuable method for diagnosing glaucoma structurally [3]. Primary open angle glaucoma is causing a progressive optic neuropathy and its development is associated with loss of tissue in the neuroretinal rim of the optic disc and that will lead to increase in the size of the optic cup. The pattern of neuroretinal rim loss and cup enlargement may take the form of focal or diffuse change, or both in combination. Focal change, with the loss of the physiological shape of the neuroretinal rim, is identified by careful clinical examination. Diffuse change, with maintenance of the physiological rim shape, is much more difficult to identify. It is in these cases that quantification of the neuroretinal rim area or cup size is useful. Methods have been described to estimate the area of the neuroretinal rim during ophthalmoscopic examination, but several measurements and calculations or additional equipment are required. Clinical estimation of the size of the cup using either the slit lamb or a simple imaging modalities such as fundus images is a significant clinical parameter and remains the simplest and most frequently performed assessment of the optic disc in the diagnosis and follows up the progression of the glaucoma suspect. The estimation of the size of the cup is usually made by comparison with the size of the disc and given as the ratio of the vertical and horizontal diameter of the cup to the vertical and horizontal diameter of the disc based on Garway-Heath et al. [4]. Thus, an automatic system for examination of optic nerve head is very useful. In a recent paper, Almazroa et al. [5] critically review the literature on glaucoma image processing.Recently Dhumane and Patil [6] have developed an algorithm for calculating the cup to disc ratio. In this algorithm superpixel segmentation was used to extract disc and boundaries. Thirty-seven images were used to test the algorithm and it successfully segmented 33 images. Guerre et al. [7] introduced a technique based on Otsu’s adaptive thresholding and a support vector machines classifier with linear kernel. The algorithm was tested on two datasets (29 and 26 images), and the accuracies of the cup to disc ratio were 89% and 59%, respectively. Zilly et al. [8] proposed a novel convolutional neural network based method for optic cup and disc segmentation. To reduce computational complexity, an entropy based sampling technique was introduced. The algorithm was tested using 10 images and the overlap was 89.5% between the segmented disc and ground truth, and 86.4% between the segmented cup and ground truth. Issac et al. [9] introduced a technique based on adaptive thresholding using features from the image such as mean and standard deviation. The algorithm was tested on 63 images and the accuracy was 92.06%. Alghmdi et al. [10] developed an automatic system to measure the cup to disc ratio based on superpixels clustering algorithm using simple linear iterative clustering and a feed-forward neural network classifier. The algorithm was tested using 60 images and the mean nonoverlapping error was 11% for the disc and 29% for the cup.This paper gives the results from calculations of the horizontal and vertical cup to disc ratios using our previously introduced optic disc [11] and cup [12] algorithms. The algorithms were tested using the RIGA dataset. The rest of the paper is organized as follows. The methodology of the research is explained in Section 2. Results are presented in Section 3. We discuss the results and conclude in Section 4.2. Methodology2.1. DatasetRIGA dataset was collected in order to facilitate research on computer-aided diagnoses of glaucoma. The dataset consists of 750 color fundus images obtained from three different resources: (1) 460 images from MESSIDOR images dataset [13] with two images of sizes pixels and pixels, (2) 195 images from Bin Rushed Ophthalmic center in Riyadh, Saudi Arabia. They were acquired in 2014 using a Canon CR2 Nonmydriatic digital retinal camera (less resolution images). The images sizes are pixels. An additional 95 images were obtained from Magrabi Eye Center in Riyadh, Saudi Arabia. The images were acquired between 2012 and 2014 using a TOPCON TRC 50DX mydriatic retinal camera (more resolution images). The images sizes are pixels. The images were notated manually by 6 ophthalmologists individually. Each one notated the disc and cup boundaries manually using a precise pen for Microsoft surface pro 3 with 12 inches high resolution screen ( pixels). Six parameters were calculated for the manual marking in order to be used to evaluate the algorithms, namely, disc area, disc centroid, cup area, cup centroid, vertical cup to disc ratio, and horizontal cup to disc ratio. The 3 datasets contain both normal and glaucomatous fundus images.The dataset was divided into two sets: training set with 200 images and testing set with 550 images for the training and testing purpose for the developed algorithms (Table 1).Table 1: Dataset information.2.2. Optic Disc and Cup SegmentationBriefly, the optic nerve head was localized using the procedures explained by Almazroa et al. [14] and Burman et al. [15] and optic disc segmentation was introduced by Almazroa et al. [11] based on inpainting the blood vessels and level set method. A fast digital image inpainting technique [16] was applied. The blood vessels were extracted; thus the extracted blood vessels are utilized to be the mask which identifies the area that wants to be inpainted. Blood vessels were extracted using a top-hat transform on the G-channel of the fundus image. In the second step, the segmentation process represented by the active contour model implemented by the level set [17] was applied. Based on the quality of the image, one of the two paths was considered for applying the level set (Figure 1). From the three sets of images in RIGA dataset, Bin Rushed images are low quality and need a double level set. After applying the first level set, the contour was considered as a second optic disc localization in order to restrict the variations from the center that cause the problems. Then the second localization was split into two to apply the level set again in order to obtain a more accurate segmentation.Figure 1: The final algorithm flowchart.On the other hand, the cup segmentation was introduced by Almazroa et al. [12]. The blood vessels were extracted using the same approach as that used for optic disc segmentation. Image thresholding was applied using an Interval Type-II fuzzy entropy based thresholding scheme with Differential Evolution on the localized image to detect the intensity of the optic cup borders. Hough transform was used to approximate the cup boundaries.Four loops were considered for four different threshold values based on some conditions which will be discussed later.The training images were the keys for developing this algorithm. The aforementioned thresholding technique was used in order to detect the cup boundaries based on the image intensity. Since the cup shape and structure are not constant among the people either for normal or for glaucomatous cases, that leads to making the boundary segmentation process more complicated. Using the thresholding technique, the image will be thresholded into binary image (black and white); that is, if threshold value (1) is applied (simple threshold concept), and three image intensities if threshold value (2) is applied and so forth. Therefore, different threshold values starting from 1 up to 30 were applied using the 200 training images in order to find out the comprehensive threshold value which will detect the cup boundaries for different images sizes and quality as well as different cup cases. The 200 training images with the six ophthalmologists manual annotations were the guide for choosing the best threshold value. Therefore, four loops for
Article 4 Reads 0 Citations A Comparison of Off-Grid-Pumped Hydro Storage and Grid-Tied Options for an IRSOFC-HAWT Power Generator Mahdi Majidniya, Kobra Gharali, Kaamran Raahemifar Published: 01 January 2017
International Journal of Rotating Machinery, doi: 10.1155/2017/4384187
DOI See at publisher website ABS Show/hide abstract
An Internal Reforming Solid Oxide Fuel Cell (IRSOFC) is modeled thermodynamically; a Horizontal Axis Wind Turbine (HAWT) is designed; the combined IRSOFC-HAWT system should create a reliable source of electricity for the demand of a village located in Manjil, Iran. The output power of HAWT is unstable, but by controlling the fuel rate for the IRSOFC it is possible to have a stable power output from the combined system. When the electricity demand is over the peak or the wind speed is low/unstable/significantly high, the generated power may not be sufficient. To solve this problem, two scenarios are considered: connecting to the grid or using a Pumped Hydro Storage (PHS). For the second scenario, the extra produced electricity is saved when the production is more than demand and can be used when the extra power is needed. The economic analysis is done based on the economic conditions in Iran. The results will show a period of return about 9.5 and 13 years with the levelized cost of electricity about 0.0747 and 0.0882 $/kWh for the first and second scenarios, respectively. Furthermore, effects of some parameters such as the electricity price and the real interest rate are discussed.1. IntroductionMaking a balance between energy productions and energy demands is a challenging issue for Renewable Energy (RE) systems. Energy production from RE sources is affected by low efficiencies, high capital costs, unreliable energy sources, and environmental impacts. Recently, combined systems are introduced to create efficient and economical power sources and mostly to overcome the unreliability of RE systems such as wind turbines. Based on the local conditions and the main energy source, the way that systems are combined varies. One of the options is combining the systems with a fuel cell. A fuel cell is a Renewable Energy source with controllable output electricity. A fuel cell has a short startup time. It is possible to add or remove cells immediately from the power cycle without any trouble. An SOFC is a common type of fuel cells for power plant applications. Recent studies include an SOFC in the power generation systems. In the following, some of the recent combinations of SOFC with other systems are discussed.Soheyli et al. [1] analyzed a combined system in a hypothetical hotel in Kermanshah, Iran. The final optimized system was a combination of 10 wind turbines, 430 PV modules, 11 SOFCs, 106 batteries, and 2 heat storage tanks. Their results showed that the fuel consumption and pollution were decreased 263 and 353 times, respectively. Fathy [2] studied a system including photovoltaic modules, wind turbines, and fuel cells in Egypt. Obara et al. [3] used a numerical analysis to stabilize photovoltaic cyclic fluctuations by using a governor-free control of an SOFC and a flywheel inertia system. Hosseini et al. [4] analyzed a combination of photovoltaic-electrolyzer, SOFC, and a heat recovery unit for a residential area. The system supported an absorption chiller and a heat recovery steam generator using the excess heat of an SOFC. They conducted an energy and exergy analysis of the system. Vigneysh and Kumarappan [5] studied a combination of photovoltaic cells, an SOFC, and a battery energy storage system. For making a balance between power generation and consumption, a fuzzy logic method was applied. The SOFC was a backup generator. It reached its rated power when the charge of the battery energy storage system was at the minimum limit. Tenfen and Finardi [6] presented a mathematical model to manage the energy from a microgrid. They modeled a micro turbine and an SOFC and combined them with a battery bank, photovoltaic cells, and a wind turbine. Akikur et al. [7] developed a mathematical model for a reversible SOFC combined with photovoltaic cells. To produce 100 kW of electricity, they investigated thermodynamically and economically three modes: a solar-SOFC mode for low solar radiation, a solar-solid oxide steam electrolyzer mode for high solar radiation, and an SOFC mode at night. Bai et al. [8] modeled and controlled an SOFC with photovoltaic cells. They used two PI controllers using an artificial bee colony.It was shown that the combination of an SOFC with a wind turbine or photovoltaic cells is a state-of-the-art topic. In this study, a PHS is also added to a combined system of a HAWT and an IRSOFC. The system can run on grid or off grid with the aid of PHS as an energy storage component. The results are compared based on the economic and thermodynamic analysis to study the feasibility of each scenario; this comparison has not been done before based on the knowledge of the authors. In addition, the systems should fulfill the power demand of a village located in Manjil, Iran. The parameters are chosen based on the local economic and environmental conditions, making this study unique. It should be noticed that the system is sized based on monthly average wind speed and the needed electricity. That means the wind speed may not always match the needed electricity. For solving this problem, the two following scenarios are introduced.The First Scenario (Figure 1(a)). The system is connected to the grid. It is possible to sell and buy electricity from the grid. The electricity can be sold to the grid when the produced electricity is more than demand. The electricity can be bought from the grid when the production cannot meet the demand. It is assumed that the prices of the selling and buying electricity are the same.Figure 1: Schematic of the system: (a) first scenario and (b) second scenario.The Second Scenario (Figure 1(b)). The system is not connected to the grid. The extra electricity production of the wind turbine is saved by a Pumped Hydro Storage (PHS) system and can be released when it is needed. A PHS is a system that contains a pump, a turbine, and two reservoirs, a higher reservoir and a lower one. During high electricity production, the PHS pumps the water from the lower reservoir to the higher one to save the energy. When the electricity is needed, the water goes from the higher reservoir to the lower one through a turbine. Because of the pump and the turbine efficiencies and the other losses, some of the energy will be lost during these processes. To make up this energy, the size of the HAWT has been increased. The size of the PHS is assumed to be equal to the maximum needed electricity minus maximum production of the IRSOFC.The case study is to produce power for a village with 100 houses located in Manjil, Iran. An economic analysis will show the possibility of the study.Based on two scenarios, the main components, the IRSOFC and the HAWT, are modeled and designed in Sections 2 and 3. The economic method is described in Section 4. After validating the IRSOFC and the HAWT in Section 5, the results of the modeling and the economic analysis are discussed in Section 6.2. Thermodynamic Modeling of IRSOFCThe modeling of components is done by home codes. The assumptions of the fuel cell modeling are(i)ideal gases,(ii)linear variation of enthalpy,(iii)working in a steady condition,(iv)being isolated with no heat exchange,(v)no contact resistance,(vi)ignoring radiation heat transfer between the gas canal and the solid container,(vii)tubular cells and stack.2.1. Direct Internal ReformerFor a fuel cell consuming natural gas, an internal or an external reformer is needed. Since an external reformer needs a cooling system and is less cost-effective than an internal one, the internal reformers are preferred [9].In the internal reforming process, the occurring reactions are very endothermic, while the heat is provided by the fuel cell. The reactions in this process are the steam reforming reaction () and the gas water shifting reaction (). According to these reactions, natural gas transforms to hydrogen in the fuel cell. Hydrogen takes part in an electrochemical reaction (). Here, , , and are the molar rate progress of reforming, shifting, and electrochemical reactions, respectively [10]. This molar rate progress is calculated based on the relation among the equation constants, the temperature, and the pressure of the gasses.2.2. Reversible VoltageThe reversible voltage () of the cell can be found by using Nernst equation:Because of the irreversibilities, the voltage of a real fuel cell is less than that of a reversible one. These irreversibilities are called overpotentials. Overpotentials can be divided into three categories: Ohmic (), activation (), and concentration (). The magnitude of these overpotentials grows as current density increases, which results in reducing the voltage of the cell. The cell voltage can be described as ([11])2.2.1. Ohmic OverpotentialOhmic overpotential () can be calculated aswhere is the resistance: is the thickness (Table 1) of components and is the resistivity:Table 1: IRSOFC resistivity constants [12]. and are resistivity constants (Table 1).2.2.2. Activation OverpotentialButler–Volmer equation [13] describes is the heat transfer coefficient and is the exchange current density. The heat transfer coefficient is the ratio of the exchange in the potential drop which results in changing the rate of constants. The value of for a fuel cell is 0.5. For , (6) can beTherefore,Equation (7) should be applied to both anode and cathode [13]. (preexponential factor) like the activation energy () depends on electrode materials [11] (Table 2).Table 2: Values of activation overpotential constants [11].2.2.3. Concentration OverpotentialConcentration overpotential can be calculated as where is the limiting current density. Considerable growth of concentration overpotential will be observed if the current density meets its limit.3. HAWT ModelingA home code for designing a HAWT is written based on Blade Element Momentum (BEM) method. For a rotor of a HAWT with the radius (), the blade should be divided into some sections or elements. Each section () has a distance from the hub of the rotor.For each element
Article 4 Reads 1 Citation Agreement among ophthalmologists in marking the optic disc and optic cup in fundus images Ahmed AlMazroa, Sami Alodhayb, Essameldin Osman, Eslam Ramad... Published: 30 August 2016
International Ophthalmology, doi: 10.1007/s10792-016-0329-x
DOI See at publisher website PubMed View at PubMed