Pusat Pengajian Kejuruteraan Mekanikal - Tesis

Browse

Now showing 1 - 5 of 262

Application of Asymptotic Waveform Evaluation (AWE) In Beam and Truss
(2004-02)
Lau, Sun Wah
There is always a misunderstanding that, when a force is applied on a structure (either beam or truss), the maximum deflection or displacement of the material is that shown in steady state. Actually, these structures will deform more than the displacement during the steady state. This is due to the moment caused by the forces as stated in Newton’s second law of motion (F=ma). Because of this unexpected higher magnitude of displacement, many products have failed to achieve its desirable quality. Micro-scale electronic packaging is a very good example. The wire boding equipment causes excessive deflection on electronic package, and damage the tiny component in the package. Earthquake, an undesirable and unexpected disaster, transferring vibration on bridge trusses. Most of the cases, the impulse force from earth fails the structure of bridge trusses. Therefore, dynamic analysis on structure is essential nowadays. Plenty of analysis procedures has been introduced. Among these methods, Finite element method (FEM) has given an accurate result besides of its flexibility. The FEM is a numerical method for solving problems of engineering and mathematical physics (Logan, 2001). However, implementation of Finite Element Method (FEM) in structural and other analysis usually will produce a formulation in space/time domain. This kind of space/time domain formulation leads to a set of ordinary differential equation and have to be solved in the time domain. An implementation of AWE scheme in first and second order ordinary differential equation shows a break through as compared with conventional method. This advanced, powerful and efficient scheme shows excellent result in electronic and thermal analysis (Ooi, 2003; Da-Guang Liu, 1995). In this thesis, AWE is pioneered in beam and truss analysis. Steady state response and dynamic response (before steady state) will be considered.
Characterization of piezoelectric patch and its application to the active vibration suppression for cantilever beam
(2019-08)
Mohd Hafiz Abdul Satar
Piezoelectric is a versatile material where it can be used either as actuator or sensor in many application due to the mechanical and electrical energy conversion. However, the piezoelectric performance is limited by its non-linear characteristics. These characteristics are widely studied but less explored especially involving all the four non-linearity effects (hysteresis, saturation, creep and uncertainty vibration). Thus, this study aims to characterize all four non-linear characteristics of the piezoelectric patch as sensor and actuator and its application to active vibration suppression (AVS) system for a cantilever beam. Then, the piezo is applied on the AVS system to study the performance in three different parameters (frequency independent, controller tuning methods and sine swept). From the study, the result shows that, the hysteresis saturation and uncertainty vibration were significantly found in the actuator and sensor characterizations, while the creep was clearly observed in actuator application. In overall, these four non-linear characteristics were getting worse as the operating frequency and input voltage were higher. At different operating frequencies of AVS, the uncertainty vibration reduction was increased linearly except at 500 Hz. The proportional gain step-up (PGS) was the best tuning method for the AVS system. The frequency dependent study also shown good performance of vibration reduction within the range of 26 Hz to 245 Hz. The main contribution of this study is the characterization of all four piezoelectric non-linear characteristics and its application on the AVS system which shown a significant vibration reduction.
Low-cost condition monitoring for unbalanced motor systems using tuned dynamic vibration absorber
(2025-02)
Mohd Affan bin Mohd Rosli
Unbalanced motor is referred to the situation of a rotating system where there is an uneven distribution of mass, resulting the significant vibration or imbalance problems. Poor management of unbalanced motor can lead to various issues, such as increased vibration, decreased efficiency, and potential damage of both motor and the integrated system. This study investigates the performance of reduction vibration for an integrated unbalanced motor-beam structure using a Tuned Dynamic Vibration Absorbers (TDVA), with different types of TDVA stiffness (stainless steel, aluminium, brass and titanium). To gain a better nderstanding of the system dynamic behavior, the natural frequencies of the beam were determined using an Experimental Modal Analysis prior to implementing the TDVA. The Operational Deflection Shapes (ODS) experiment was conducted in the z-axis direction with three different motor speeds; 880 RPM (14.8 Hz), 2100 RPM (35 Hz) and 2800 RPM (46.5 Hz) to observe the most significant vibration of the beam during operation. Later, the TDVA which consisted of two secondary masses, was employed to modify the structural dynamic response of the beam. The lengths of the TDVA masses were adjusted based on the motor speed to optimize vibration reduction of the beam. The selection of TDVA stiffness materials was driven by their varying densities, moduli of elasticity and damping capacities, providing insight into their suitability for specific operating frequencies and conditions. Various TDVA stiffness materials were applied to determine the most effective vibration attenuation and it was found that aluminium material has produced the highest attenuation of 93.18 % at motor speed of 2880 RPM. Furthermore, a low-cost condition-based monitoring (CBM) system was developed using an Arduino Uno microcontroller connected to a Raspberry Pi. This system utilized an MPU9250 sensor which is cost-effective and appropriate for vibration measurement. The CBM system dashboard was hosted using the cloud, allowing real-time access to the vibration data. The system employed four programmable conditions to continuously assess the vibration activities. This affordable approach offers an accessible solution for small-scale industries, reducing reliance on expensive industrial-grade analyzers. In the event of abnormal vibration, the CBM system can trigger a notification alert, serving as a preventive measure against structure failures. The findings contribute to broader applications, including the improvement of maintenance strategies across various industries, emphasizing the transformative impact of combining effective vibration control and low-cost monitoring systems. It is also contributes to the understanding of the effect of different TDVA stiffness materials on the vibration control of beam structures with the additional of practical approach for real-time condition monitoring to improve system reliability.
Performances Of Metaheuristic Algorithms In Optimizing Tool Capacity Allocations
(2014-05)
Goheannee
Semiconductor manufacturing industry in general has moved into high mix productions resulting from the drastic pace of product innovation. Capacity planning In semiconductor manufacturing facility, such as allocating right mix of products to maximize the capacity output, needs to consider multiple mutually influenced constraints in resource, product demand, as well as product and process characteristics. To achieve the best allocation, optimization methods, such as metaheuristic algorithms are commonly used. This research compares the performances of various metaheuristic algorithms to optimize tool capacity allocation in two case studies. In this research, the algorithms studied includes Genetic Algorithm, Particle Swarm Optimization Algorithm, Differential Evolution Algorithm, Harmony Search Algorithm, Teaching-LearningBased Optimization Algorithm and Black Hole Algorithm. These algorithms are inspired by different nature of phenomenon. The former three are common in literature for tool capacity allocation problems. The latter three are the next generation of metaheuristic algorithms and albeit popular elsewhere, have no known attempt in tool capacity allocation problems. The case studies were obtained from two real industries and five demand scenarios were derived. The demand scenarios were with different demand intensities and levels. For each case study, a capacity model was constructed in Microsoft Excel spreadsheet, as an input to the above mentioned metaheuristic algorithms which programmed in Matlab coding. The performances considered are tool utilization and aggregate capacity outputs.
Ginger seed growth recognition using mask region based convolutional neural network (mask r-cnn)
(2023-01-01)
Tong Yin Syuen
As a plant that poses unique culinary and medical uses, ginger has emerged as a valuable commodity in Asia. Among the critical processes in the production of ginger is ginger seed preparation. It is particularly important to monitor the growth and quality of ginger seeds before they are being sown in growing media to ensure germination. However, to date, the ginger seed monitoring process remains manual and is reliant on human experts, despite the growing demand for more effective and accurate monitoring. In this work, a total 1,746 images consisting 2,230 sprout instances were collected from 282 ginger seed samples. In order to realize the automatic monitoring of ginger seeds, deep learning architectures were employed to detect the ginger seed sprouts in three stages from the digital images. This work assessed and compared the instance segmentation task using end-to-end Mask R-CNN models built by different strategies. Then, a two-stage hybrid detector-classifier model was also proposed to benefit from model task specialization concept. Specifically, an end-to-end binaryclass Mask R-CNN and multi-class classifier were combined to be compared to an end-to-end multi-class Mask R-CNN. The experimental results indicate that the use of the hybrid detector-classifier model developed in this work achieved mAP0.50 of 84.27% at inference time of 0.383 second per image in the detection of 402 images consisting of 514 sprout instances. Besides, substantial confusion between object classes in the model was also observed to be in line with the human expert’s perception in data annotation.

Browse

Recent Submissions