CERES :: Browsing by Author "Rana, Zeeshan A."

Browsing by Author "Rana, Zeeshan A."

Now showing 1 - 20 of 30

Open Access
A 3D CFD analysis of flow past a hipped roof with comparison to industrial building standards
(Techno Press, 2022-06-25) Khalil, Khalid; Khan, Huzafa; Chahar, Divyansh; Townsend, Jamie F.; Rana, Zeeshan A.
Three-dimensional (3D) computational fluid dynamics (CFD) analysis of flow around a hipped-roof building representative of UK inland conditions are conducted. Unsteady simulations are performed using three variations of the k-ϵ RANS turbulence model namely, the Standard, Realizable, and RNG models, and their predictive capability is measured against current European building standards. External pressure coefficients and wind loading are found through the BS 6399-2:1997 standard (obsolete) and the current European standards (BS EN 1991-1-4:2005 and A1:20101). The current European standard provides a more conservative wind loading estimate compared to its predecessor and the k-ϵ RNG model falls within 15% of the value predicted by the current standard. Surface shear stream-traces and Q-criterion were used to analyze the flow physics for each model. The RNG model predicts immediate flow separation leading to the creation of vortical structures on the hipped-roof along with a larger separation region. It is observed that the Realizable model predicts the side vortex to be a result of both the horseshoe vortex and the flow deflected off it. These model-specific aerodynamic features present the most disparity between building standards at leeward roof locations. Finally, pedestrian comfort and safety criteria are studied where the k-ϵ Standard model predicts the most ideal pedestrian conditions and the Realizable model yields the most conservative levels.
Open Access
Aerodynamic and structural design of a 2022 Formula One front wing assembly
(MDPI, 2020-12-09) Castro, Xabier; Rana, Zeeshan A.
The aerodynamic loads generated in a wing are critical in its structural design. When multi-element wings with wingtip devices are selected, it is essential to identify and to quantify their structural behaviour to avoid undesirable deformations which degrade the aerodynamic performance. This research investigates these questions using numerical methods (Computational Fluid Dynamics and Finite Elements Analysis), employing exhaustive validation methods to ensure the accuracy of the results and to assess their uncertainty. Firstly, a thorough investigation of four baseline configurations is carried out, employing Reynolds Averaged Navier–Stokes equations and the k-ω SST (Shear Stress Transport) turbulence model to analyse and quantify the most important aerodynamic and structural parameters. Several structural configurations are analysed, including different materials (metal alloys and two designed fibre-reinforced composites). A 2022 front wing is designed based on a bidimensional three-element wing adapted to the 2022 FIA Formula One regulations and its structural components are selected based on a sensitivity analysis of the previous results. The outcome is a high-rigidity-weight wing which satisfies the technical regulations and lies under the maximum deformation established before the analysis. Additionally, the superposition principle is proven to be an excellent method to carry out high-performance structural designs.
Open Access
Classification of flow regimes using a neural network and a non-invasive ultrasonic sensor in an S-shaped pipeline-riser system
(Elsevier, 2021-11-24) Nnabuife, Somtochukwu Godfrey; Kuang, Boyu; Rana, Zeeshan A.; Whidborne, James F.
A method for classifying flow regimes is proposed that employs a neural network with inputs of extracted features from Doppler ultrasonic signals of flows using either the Discrete Wavelet Transform (DWT) or the Power Spectral Density (PSD). The flow regimes are classified into four types: annular, churn, slug, and bubbly flow regimes. The neural network used in this work is a feedforward network with 20 hidden neurons. The network comprises four output neurons, each of which corresponds to the target vector's element number. 13 and 40 inputs are used for features extracted from PSD and DWT respectively. Experimental data were collected from an industrial-scale multiphase flow facility. Using the PSD features, the neural network classifier misclassified 3 out of 31 test datasets in the classification and gave 90.3% accuracy, while only one dataset was misclassified with the DWT features, yielding an accuracy of 95.8%, thus showing the superiority of the DWT in feature extraction of flow regime classification. The approach demonstrates the applicability of a neural network and DWT for flow regime classification in industrial applications using a clamp-on Doppler ultrasonic sensor. The scheme has significant advantages over other techniques as only a non-radioactive and non-intrusive sensor is used. To the best of our knowledge, this is the first known successful attempt for the classification of liquid-gas flow regimes in an S-shape riser system using an ultrasonic sensor, PSD-DWTs features, and a neural network.
Open Access
Computational aerodynamics analysis of non-symmetric multi-element wing in ground effect with humpback whale flipper tubercles
(MDPI, 2020-12-17) Arrondeau, Benjamin; Rana, Zeeshan A.
The humpback whale flipper tubercles have been shown to improve the aerodynamic coefficients of a wing, especially in stall conditions, where the flow is almost fully detached. In this work, these tubercles were implemented on a F1 front-wing geometry, very close to a Tyrrell wing. Numerical simulations were carried out employing the k−ω SST turbulence model and the overall effects of the tubercles on the flow behavior were analyzed. The optimal amplitude and number of tubercles was determined in this study for this front wing where an improvement of 22.6% and 9.4% is achieved, respectively, on the lift and the L/D ratio. On the main element, the stall was delayed by 167.7%. On the flap, the flow is either fully detached, in the large circulation zone, or fully attached. Overall, in stall conditions, tubercles improve the downforce generation but at the cost of increased drag. Furthermore, as the tubercles are case-dependent, an optimal configuration for tubercles implementation also exists for any geometry.
Open Access
Computational engineering analysis of external geometrical modifications on the MQ-1 unmanned combat aerial vehicle
(Elsevier, 2020-03-17) Bagul, Prakash; Rana, Zeeshan A.; Jenkins, Karl W.; Könözsy, László Z.
This paper focuses on the effects of external geometrical modifications on the aerodynamic characteristics of the MQ-1 predator Unmanned Combat Aerial Vehicle (UCAV) using computational fluid dynamics. The investigations are performed for 16 flight conditions at an altitude of 7.6 km and at a constant speed of 56.32 m/s. Two models are analysed, namely the baseline model and the model with external geometrical modifications installed on it. Both the models are investigated for various angles of attack from −4° to 16°, angles of bank from 0° to 6° and angles of yaw from 0° to 4°. Due to the unavailability of any experimental (wind tunnel or flight test) data for this UCAV in the literature, a thorough verification of calculations process is presented to demonstrate confidence level in the numerical simulations. The analysis quantifies the loss of lift and increase in drag for the modified version of the MQ-1 predator UCAV along with the identification of stall conditions. Local improvement (in drag) of up to 96% has been obtained by relocating external modifications, whereas global drag force reduction of roughly 0.5% is observed. The effects of external geometrical modifications on the control surfaces indicate the blanking phenomenon and reduction in forces on the control surfaces that can reduce the aerodynamic performance of the UCAV.
Open Access
Constraints on optimising encoder-only transformers for modelling sign language with human pose estimation keypoint data
(MDPI, 2023-11-02) Woods, Luke T.; Rana, Zeeshan A.
Supervised deep learning models can be optimised by applying regularisation techniques to reduce overfitting, which can prove difficult when fine tuning the associated hyperparameters. Not all hyperparameters are equal, and understanding the effect each hyperparameter and regularisation technique has on the performance of a given model is of paramount importance in research. We present the first comprehensive, large-scale ablation study for an encoder-only transformer to model sign language using the improved Word-level American Sign Language dataset (WLASL-alt) and human pose estimation keypoint data, with a view to put constraints on the potential to optimise the task. We measure the impact a range of model parameter regularisation and data augmentation techniques have on sign classification accuracy. We demonstrate that within the quoted uncertainties, other than ℓ2 parameter regularisation, none of the regularisation techniques we employ have an appreciable positive impact on performance, which we find to be in contradiction to results reported by other similar, albeit smaller scale, studies. We also demonstrate that the model architecture is bounded by the small dataset size for this task over finding an appropriate set of model parameter regularisation and common or basic dataset augmentation techniques. Furthermore, using the base model configuration, we report a new maximum top-1 classification accuracy of 84% on 100 signs, thereby improving on the previous benchmark result for this model architecture and dataset.
Open Access
Development of gas-liquid flow regimes identification using a noninvasive ultrasonic sensor, belt-shape features, and convolutional neural network in an S-shaped riser
(IEEE, 2021-07-14) Nnabuife, Somtochukwu Godfrey; Kuang, Boyu; Whidborne, James F.; Rana, Zeeshan A.
The problem of classifying gas-liquid two-phase flow regimes from ultrasonic signals is considered. A new method, belt-shaped features (BSFs), is proposed for performing feature extraction on the preprocessed data. A convolutional neural network (CNN/ConvNet)-based classifier is then applied to categorize into one of the four flow regimes: 1) annular; 2) churn; 3) slug; or 4) bubbly. The proposed ConvNet classifier includes multiple stages of convolution and pooling layers, which both decrease the dimension and learn the classification features. Using experimental data collected from an industrial-scale multiphase flow facility, the proposed ConvNet classifier achieved 97.40%, 94.57%, and 94.94% accuracy, respectively, for the training set, testing set, and validation set. These results demonstrate the applicability of the BSF features and the ConvNet classifier for flow regime classification in industrial applications.
Open Access
Development of vision guided real-time trajectory planning system for autonomous ground refuelling operations using hybrid dataset
(AIAA, 2023-01-19) Yildirim, Suleyman; Rana, Zeeshan A.; Tang, Gibert
Accurate and rapid object localisation and pose estimation are playing key roles during some of the real-time robotic operations such as object grasping and object manipulating. To do so, high-level robotic vision solutions need to be adopted. Computer vision approaches require a large amount of data to be able to create a perception pipeline robustly. Preparing such dataset to train the deep neural network could be challenging as the collection and manual annotation of huge amounts of data can take long hours and the development of the dataset needs to cover different conditions in weather and lighting. To ease this process, generating a synthetic dataset could be used. Due to the limitations of the synthetic dataset which will be described further down, instead of using a sole synthetic dataset, a hybrid dataset can be developed with the real dataset to overcome the limitations of both datasets. Even though the main objective of this study is to fulfil an autonomous nozzle insertion process for the ground refuelling operation of civil aircraft, the proposed approach is generic and can be adapted to any 3D visual robotic manipulation operation. This study is also offered to be the first visual trajectory planning control mechanism depending on the hybrid dataset to this date.
Open Access
Drone model classification using convolutional neural network trained on synthetic data
(MDPI, 2022-08-12) Wisniewski, Mariusz; Rana, Zeeshan A.; Petrunin, Ivan
We present a convolutional neural network (CNN) that identifies drone models in real-life videos. The neural network is trained on synthetic images and tested on a real-life dataset of drone videos. To create the training and validation datasets, we show a method of generating synthetic drone images. Domain randomization is used to vary the simulation parameters such as model textures, background images, and orientation. Three common drone models are classified: DJI Phantom, DJI Mavic, and DJI Inspire. To test the performance of the neural network model, Anti-UAV, a real-life dataset of flying drones is used. The proposed method reduces the time-cost associated with manually labelling drones, and we prove that it is transferable to real-life videos. The CNN achieves an overall accuracy of 92.4%, a precision of 88.8%, a recall of 88.6%, and an f1 score of 88.7% when tested on the real-life dataset.
Open Access
Drone model identification by convolutional neural network from video stream
(IEEE, 2021-11-15) Wisniewski, Mariusz; Rana, Zeeshan A.; Petrunin, Ivan
We present a convolutional neural network model that correctly identifies drone models in real-life video streams of flying drones. To achieve this, we show a method of generating synthetic drone images. To create a diverse dataset, the simulation parameters (such as drone textures, lighting, and orientation) are randomized. This synthetic dataset is used to train a convolutional neural network to identify the drone model: DJI Phantom, DJI Mavic, or DJI Inspire. The model is then tested on a real-life Anti-UAV dataset of flying drones. The benchmark results show that the DenseNet201 architecture performed the best. Adding Gaussian noise to the training dataset and performing full training (as opposed to freezing layers) shows the best results. The model shows an average accuracy of 92.4%, and an average precision of 88.6% on the test dataset.
Open Access
Enhancing aircraft safety through advanced engine health monitoring with long short-term memory
(MDPI, 2024-01-14) Yildirim, Suleyman; Rana, Zeeshan A.
Predictive maintenance holds a crucial role in various industries such as the automotive, aviation and factory automation industries when it comes to expensive engine upkeep. Predicting engine maintenance intervals is vital for devising effective business management strategies, enhancing occupational safety and optimising efficiency. To achieve predictive maintenance, engine sensor data are harnessed to assess the wear and tear of engines. In this research, a Long Short-Term Memory (LSTM) architecture was employed to forecast the remaining lifespan of aircraft engines. The LSTM model was evaluated using the NASA Turbofan Engine Corruption Simulation dataset and its performance was benchmarked against alternative methodologies. The results of these applications demonstrated exceptional outcomes, with the LSTM model achieving the highest classification accuracy at 98.916% and the lowest mean average absolute error at 1.284%.
Open Access
Evaluation of the SU2 open-source code for a hypersonic flow at mach number 5
(Miskolci Egyetemi Kiadó, 2022-11-10) Yeap, Jia-Ming; Rana, Zeeshan A.; Könözsy, László Z.; Jenkins, Karl W.
This paper presents the evaluation of the Stanford University Unstructured (SU2) open-source computational software package for a high Mach number 5 flow. The test case selected is an impinging shock wave turbulent boundary layer interaction (SWTBLI) on a flat plate where the experimental data of Sch¨ulein et al. [27] is used for validation purposes. Two turbulence models, the Spalart–Allmaras (SA) and the k-ω Shear Stress Transport (SST) within the SU2 code are evaluated in this study. Flow parameters, such as skin friction, wall pressure distribution and boundary layer profiles are compared with experimental values. The results demonstrate the performance of the SU2 code at a high Mach number flow and highlight its limitations in predicting fluid flow physics. At higher shock generator angles, the discrepancy between experimental and CFD data is more significant. Within the interaction and flow separation zones, a smaller separation bubble and delayed separation are predicted by the SA model while the k-ω SST model predicts early separation. Both models are able to predict wall pressure distribution correctly within the experimental values. However, discrepancies were observed in the prediction of skin friction due to the inability of the models to capture the boundary layer recovery after shock impingement.
Open Access
Gas-liquid flow regimes identification using non-intrusive Doppler ultrasonic sensor and convolutional recurrent neural networks in an S-shaped riser
(Elsevier, 2022-01-19) Kuang, Boyu; Nnabuife, Somtochukwu Godfrey; Sun, Shuang; Whidborne, James F.; Rana, Zeeshan A.
The problem of gas-liquid (two-phase) flow regime identification in an S-shaped riser using an ultrasonic sensor and convolutional recurrent neural networks (CRNN) is addressed. This research systematically evaluates three different schemes with four CRNN-based classifiers over fourteen experiments. Four metrics are used as the evaluation criteria: categorical accuracy, categorical cross-entropy, mean square error (MSE), and computation graph complexity. Compared with existing results, a compatible performance is achieved while considerably reducing the model complexity. The testing and validation accuracies were 98.13% and 98.06%, while the complexity decreased by 98.4% (only 117,702 parameters). The proposed approach is i) accurate, low complexity, and non-intrusive and hence suitable for industry, and ii) could provide a benchmark for flow regime identification.
Open Access
Hypersonic boundary layer reduction with optimisation of the Hyshot II intake using numerical methods
(AIAA, 2024-01-04) Burrows, Sam; Rana, Zeeshan A.; Prince, Simon A.
The Hyshot project demonstrated that the quality of flow in a scramjets combustion chamber is essential to achieve critical mass flow condition in a small profiled supersonic combustion chamber. This investigation utilises a steady RANS and real gas thermal model to explore boundary layer growth over various comparable compression ramp profiles in a hypersonic freestream condition. The results in this investigation suggested that given a constant net flow deflection angle and ramp length, the displacement thickness developing on a hypersonic compression ramp is inversely proportional to the magnitude of static pressure the ramp produces at the wall, where the highest pressures theoretically achievable are that of an isentropic compression process. A quasi isentropic compression distribution was achieved when the rate of deflection along a ramp’s length is linear. The data collected also suggests that the difference in displacement thickness growth of the quasi isentropic ramp relative to the nominal flat plate increases as the flight Reynold’s number increases, where the displacement thicknesses has the potential to reduce by over 50% relative to an equivalent flat plate. Low Reynold’s numbers have found to yield the opposite effect, where isentropic turning of a hypersonic flow may increase the difference in these cases. Such reductions have been shown to theoretically allow for a critical mass flow rate in a two dimensional scramjet combustion chamber without requiring slot bleed systems, whilst achieving the required flow conditions for the auto-ignition of hydrogen. It is suggested that a scramjet design based upon these findings would still require some form of porous bleed system to compensate for the growth in the boundary layer as it expands about the compression ramp into the combustion chamber, which has potential to be incorporated onto an axisymetric design to optimise the effective mass flow rate per body diameter.
Open Access
Implicit and conventional large eddy simulation of flow around a circular cylinder at Reynolds number of 3900
(AIAA, 2024-01-04) Li, Zhuoneng; Da Ronch, Andrea; Rana, Zeeshan A.; Jenkins, Karl W.
The implicit Large Eddy Simulation (iLES) incorporating an unstructured 3rd-order Weighted Essential Non-Oscillatory (WENO) reconstruction method and the conventional Large Eddy Simulation with Wall Adapting Local Eddy-Viscosity (WALE) are investigated on the flow around a circular cylinder at a Reynolds number of 3900. Simulations are carried out in the framework of open-source package OpenFOAM with a 2nd-order Euler implicit time integration and Pressure-Implicit Splitting-Operator (PISO) algorithm is used for the pressure-velocity coupling. The results are compared to the high fidelity experiment and DNS data, and demonstrated a favourable performance for iLES with a 3rd-order WENO scheme on the instantaneous flow structure. The conventional LES on the prediction of mean surface pressure coefficient and velocity profiles on the wake can be beneficial by reducing the effect of Rhie-Chow interpolation. The spectral analysis reveals that the current simulations are also capturing Von Karman shedding frequencies and shear layer frequencies. Finally, distinct features of iLES and LES are discussed.
Open Access
Implicit large eddy simulation of the flow past NACA0012 aerofoil at a Reynolds number of 1x10^5
(AIAA, 2024-01-04) Li, Zhuoneng; Rana, Zeeshan A.
In this paper, the implicit Large Eddy Simulation (iLES) incorporating an unstructured 3rd-order Weighted Essential Non-Oscillatory (WENO) reconstruction method is investigated on the flow past NACA0012 aerofoil at a Reynolds number of 1 × 10^5. The flow features involve laminar separation, transition to turbulent and re-attachment. Simulations are carried out in the framework of open-source package OpenFOAM with a 2nd-order Euler implicit time integration and Pressure-Implicit Splitting-Operator (PISO) algorithm is used for the pressure-velocity coupling. Conventional LES with Wall Adapting Local Eddy Viscosity (WALE) model is also carried out as a baseline. The results are compared with Direct Numerical Simulations (DNS) under the same flow configurations. The mean quantities such as pressure coefficient and the re-attached turbulent velocity profiles are in excellent agreement with the DNS reference. On the other hand, in the transitional region, the thickness of separation bubble obtained by both iLES and LES is thinner than the DNS. The current iLES approach has achieved a 35% reduction of mesh resolution compared to wall resolving LES and 70% reduction compared to DNS, while the accuracy is mostly satisfied.
Open Access
Implicit large eddy simulations of turbulent flow around a square cylinder at Re=22,000
(Elsevier, 2021-05-07) Zeng, Kai; Li, Zhuoneng; Rana, Zeeshan A.; Jenkins, Karl W.
In this paper, the Implicit Large-Eddy Simulation (ILES) is investigated on the flow around a square cylinder incorporating an unstructured Weighted Essential Non-Oscillatory (WENO) reconstruction method for a Reynolds number of 22,000. Simulations are undertaken in the framework of open-source package OpenFOAM and additional implicit 2nd/3rd-order WENO scheme on the convective term of the viscous incompressible Navier-Stokes Equations. A 2nd-order Euler implicit time integration and Pressure-Implicit Splitting-Operator (PISO) algorithm is used to for the pressure-velocity coupling. Conventional LES with Wall Adapting Local Eddy Viscosity (WALE) model is also carried out as a baseline. The results are compared to high fidelity experiment, DNS data and conventional LES with dynamic Smagorinsky model from previous work. Results show favorable performance for ILES with 3rd-order WENO scheme compared with the conventional LES with dynamic Smagorinsky model and similar performance against LES with WALE model. Results also show acceptable predictions over time-averaged statistics with less computational effort for the ILES of 2nd-order WENO scheme. Shear layer flow analysis suggests that both ILES and LES face similar challenges with small quantities, such as shear stress. Finally, simulations are capturing Von Krmn vortex, Kelvin-Helmholtz instability and induced frequency changes.
Open Access
The influence of micro-expressions on deception detection
(Springer, 2023-03-16) Yildirim, Suleyman; Chimeumanu, Meshack Sandra; Rana, Zeeshan A.
Facial micro-expressions are universal symbols of emotions that provide cohesion to interpersonal communication. At the same time, the changes in micro-expressions are considered to be the most important hints in the psychology of emotion. Furthermore, analysis and recognition of these micro-expressions have pervaded in various areas such as security and psychology. In security-related matters, micro-expressions are widely used to detect deception. In this research, a deep learning model that interprets the changes in the face into meaningful information has been trained using The Facial Expression Recognition 2013 dataset. Necessary data is also obtained through live stream or video stream by detecting via computer vision and evaluating with the trained model. Finally, the data obtained is transformed into graphic and interpreted to determine whether the people are trying to deceive or not. The deception classification accuracy of the custom trained model is 74.17% and the detection of the face with high precision using the computer vision methods increased the accuracy of the obtained data and provided it to be interpreted correctly. In this respect, the study differs from other studies using the same dataset. In addition, it is aimed to facilitate the deception detection which is performed in a complex and expensive way, by making it simple and understandable.
Open Access
Integrating GRU with a Kalman filter to enhance visual inertial odometry performance in complex environments
(MDPI, 2023-10-29) Tabassum, Tarafder Elmi; Xu, Zhengjia; Petrunin, Ivan; Rana, Zeeshan A.
To enhance system reliability and mitigate the vulnerabilities of the Global Navigation Satellite Systems (GNSS), it is common to fuse the Inertial Measurement Unit (IMU) and visual sensors with the GNSS receiver in the navigation system design, effectively enabling compensations with absolute positions and reducing data gaps. To address the shortcomings of a traditional Kalman Filter (KF), such as sensor errors, an imperfect non-linear system model, and KF estimation errors, a GRU-aided ESKF architecture is proposed to enhance the positioning performance. This study conducts Failure Mode and Effect Analysis (FMEA) to prioritize and identify the potential faults in the urban environment, facilitating the design of improved fault-tolerant system architecture. The identified primary fault events are data association errors and navigation environment errors during fault conditions of feature mismatch, especially in the presence of multiple failure modes. A hybrid federated navigation system architecture is employed using a Gated Recurrent Unit (GRU) to predict state increments for updating the state vector in the Error Estate Kalman Filter (ESKF) measurement step. The proposed algorithm’s performance is evaluated in a simulation environment in MATLAB under multiple visually degraded conditions. Comparative results provide evidence that the GRU-aided ESKF outperforms standard ESKF and state-of-the-art solutions like VINS-Mono, End-to-End VIO, and Self-Supervised VIO, exhibiting accuracy improvement in complex environments in terms of root mean square errors (RMSEs) and maximum errors.
Open Access
Modelling sign language with encoder-only transformers and human pose estimation keypoint data
(MDPI, 2023-05-01) Woods, Luke T.; Rana, Zeeshan A.
We present a study on modelling American Sign Language (ASL) with encoder-only transformers and human pose estimation keypoint data. Using an enhanced version of the publicly available Word-level ASL (WLASL) dataset, and a novel normalisation technique based on signer body size, we show the impact model architecture has on accurately classifying sets of 10, 50, 100, and 300 isolated, dynamic signs using two-dimensional keypoint coordinates only. We demonstrate the importance of running and reporting results from repeated experiments to describe and evaluate model performance. We include descriptions of the algorithms used to normalise the data and generate the train, validation, and test data splits. We report top-1, top-5, and top-10 accuracy results, evaluated with two separate model checkpoint metrics based on validation accuracy and loss. We find models with fewer than 100k learnable parameters can achieve high accuracy on reduced vocabulary datasets, paving the way for lightweight consumer hardware to perform tasks that are traditionally resource-intensive, requiring expensive, high-end equipment. We achieve top-1, top-5, and top-10 accuracies of 97%, 100%, and 100%, respectively, on a vocabulary size of 10 signs; 87%, 97%, and 98% on 50 signs; 83%, 96%, and 97% on 100 signs; and 71%, 90%, and 94% on 300 signs, thereby setting a new benchmark for this task.