|
Industrial Automation Laboratory Department of Mechanical Engineering University of British Columbia |
|
Research Projects |
|
Modification and Application of Particle Swarm Optimization in Electronic Heat Sink Design Mohammed Alrasheed Ph.D. Candidate Supervisors: Dr. C. W. de Silva and Dr. M. Gadala The goal of my research is to develop and evaluate practical and efficient methodology for optimal design of the thermal system of a common electronic component. In view of its appropriateness, the concept of particle swarm optimization (PSO) is modified and adapted for this purpose. PSO is a population based stochastic optimization technique. This approach is known to provide global optimization solutions to rather complex and nonlinear problems. The approach however needs modification when using to optimize a specific problem such as the optimal design of a thermal system. Otherwise the procedure, in view of its generality, will be neither efficient nor effective. In the present investigation, we develop the means to adapt the PSO method for use in the design optimization of a heat sink. This is the primary focus of the proposed research. The main objectives of proposed research are: 1. Adapt the particle swarm optimization (PSO) method for the optimal design of the cooling system of common electronic devices, for effective cooling under optimized thermal efficiency. 2. Carry out a performance analysis and a comparative study between the developed approach and conventional and classical optimization methods used in heat sink design. 3. Develop further extensions of performance enhancement strategies for the PSO method, through benchmark simulations. Carry out a comprehensive case study of the heat sink design for a common electronic device using the developed methodology. Multi-Robot Cooperative Parts Transportation for Assembly Using Artificial Immune System Approach M. Tahir Khan Ph.D. Candidate Supervisor: Dr. Clarence W. de Silva With the rapid progress of robotic technology it is becoming increasingly common to have multiple robots working together for applications such as material transport and cooperative assembly. It is generally known that each robot has a limited intellectual power, but a robot can behave more intellectually in a group because they can interact with each other and learn by operations. Cooperative robotics is desirable for a number of reasons. First, many robotic applications are inherently distributed in space, time, or functionality, thus requiring a distributed solution. Second, it is quite possible that many applications would be solved much faster if the mission could be divided across the number of robots operating in parallel. Third, it may actually be much cheaper and more practical in many applications to build a number of less capable robots that can work together for a mission rather than trying to build one sophisticated robot which can perform the entire mission with adequate reliability. My research is also focused on developing an Immune Control Framework for Multi Robot Cooperative Part Transportation for Assembly purpose having the following capabilities: Decomposition of the task into subtasks Recognition of different parts Obstacle avoidance Self-deterministic cooperation Adaptation Use of genetic algorithms for task optimization Heterogeneous robots will accomplish the overall task. Knowledge-based Operation Optimization of Kinetic Structures Using Engineering and Architectural Knowledge Madalina Wierzbicki M.A.Sc. Candidate Supervisor: Dr. Clarence W. de Silva My research concerns the study and development of new kinetic structure topologies. The original concept of a foldable structure that I developed was based on kinematic linkages. It became the basis for further development of functional programming for addressing various applications such as shelters, exploration stations, exhibition rooms and commercial spaces. Kinetic and adaptive structures offer the possibility of adjusting spatial and functional attributes on demand. They facilitate quick deployment and reduce environmental footprint. They offer means of meeting the unprecedented functional, comfort and safety demands of modern society. I am pursuing my initial notion that the challenges of designing kinetic structures which stem from their geometrically parallel nature can be addressed by application of fuzzy logic-based optimization algorithms. My experience in hi-tech industrial design projects, familiarity with engineering design workflow as well as extensive background in architectural design and its technological and ecological concerns provide me with valuable perspective and knowledge to carry out this interdisciplinary research, which deals with developing novel concepts of kinetic and adaptive structures. Remote Monitoring and Fault Diagnosis of an Industrial Machine through Sensor Fusion Roland Haoxiang Lang M.A.Sc. Candidate Supervisor: Dr. Clarence W. de Silva My area of research is sensor fusion and machine vision with application to remote monitoring and fault diagnosis of industrial processes. The integration of information and knowledge from multiple sensors is known as data fusion. In factories and process plants, experienced engineers are able to observe multiple machinery and equipment, analyze conditions and make important operational decisions. Current computer-based monitoring and supervisory control systems have limited capabilities to carry out what is naturally possible for experienced and skilled engineers. In my research, I develop techniques and computer-based systems that can monitor, analyze, understand, and integrate information from different sensors and make “intelligent” decisions using soft computing. Humans use vision in their daily interactions with the environment. Vision sensors are becoming powerful and versatile in engineering applications. In machine vision, I apply feature based object tracking for engineering problems. In particular, I am developing a remote fault diagnosis system for an automated fish cutting machine through sensor fusion. Sound, vibration and vision sensors are used for information acquisition. A neural-fuzzy architecture is employed to fuse these three types of sensory information. After off-line training, the system is applied on-line for fault diagnosis in the fish cutting machine. Image-based Visual Servoing with Hybrid Camera Configuration for Robust Grasping Guan-Lu Zhang MASc Candidate Supervisor: Dr. C.W. de Silva Image-based visual servoing (IBVS) is better than position-based visual servoing (PBVS) because its control inputs are defined by errors in the image space. Thus, camera calibration errors and coarse modeling of the robot, which prove to be main problems for PBVS, are not issues for IBVS. However, IBVS is not free of shortcomings. The interaction matrix defines the fundamental relationship between the feature point velocities in the image and the camera velocity in 3D space. As a key parameter of the interaction matrix translational component, the unknown depth variable for each feature point at each iteration of the control loop has significant influence on the scheme’s stability, the camera’s realizable motion and the global convergence properties of image errors. The ultimate goal of my work is to improve the performance and robustness of the classical image-based visual servoing by utilizing a hybrid camera configuration. As an application to mobile manipulation in unstructured environments, a mobile robot will be augmented with a robotic arm with an eye-in-hand monocular camera and an eye-to-object stationary stereo camera. The expected output of my work can be summarized by the following Objectives 1. Improve upon the robustness of classical image-based visual servoing with depth estimation from a stationary stereo camera and an eye-in-hand monocular camera. 2. Implement soft computing techniques on maintaining target object within the camera field-of-view 3. Determine human-like grasping positions of robotic grippers using neural networks Monitoring, fault detection, and diagnosis of an industrial plant Mr. J. Ramon Campos M.ENG Candidate Supervisor: Dr. Clarence W. de Silva Engineering is high-speed traveling to an unsuspected future; achievements in science and technology allow developed countries maintain the highest standards in quality, research and progress overall. Along with technical matters, ethics and morality are important virtues in society. Taking this in consideration, my enthusiasm and interest in contributing with Mexico’s sustainable development, and the continuous increased professional competitiveness, I have decided to continue my preparation at UBC, specifically at the Industrial Automation Laboratory. Therefore, my interest is to have a closer understanding of the technological advances nowadays and combine my extensive experience in industrial product design projects as well as my background in project management and its technological and social concerns, in order to provide me with a valuable perspective and knowledge to integrate both areas as a synergetic integration, Mechatronic approach, for any project. Condition Monitoring of Industrial Machine Using Unscented Kalman Filter Mr. Behnam Razavi MASc Candidate Supervisor: Dr. C.W. de Silva The early detection of fault occurrence is critical in avoiding product deterioration, performance degradation, major damage to the machinery itself and damage to human health or even loss of lives. The quick and correct diagnosis of the faulty component facilitates the decisions on corrective actions or on repairs Fault detection and diagnosis approach for automated systems such as hydraulic systems and ongoing developments in this area is one of the main focuses of researchers in engineering. A successful performance of real-time fault detection and diagnosis in large and complex hydraulic systems is seldom achieved and presented in recent years. However, the lack of an effective method for handling temporal data is seen as one of the key problem in this area. Thus the main objective of this thesis is to select and to bring forward the best modeling approach with as much accuracy as possible in order to detect and diagnose the possible faults occurring in a complex hydraulic system in real time applications for on-line system monitoring. The goals of this approach are to achieve an effective system response for this integrated complex hydraulic system followed by developing and verifying the control system components in a best possible way. The industrial hydraulic manipulator chosen in this research for experimental investigation of the issues related to fault detection and isolation (FDI) is a planar electro- hydraulic actuation system which is an integral part of an automated fish processing machine. A condition monitoring technique based on the Unscented Kalman Filter (UKF) is considered in this project on a validated model. The UKF algorithm estimates the system states and generates the residual errors. The faults are investigated in this work are as followed: 1)the external chamber leakage at either side of the actuator, 2) the internal leakage between the two hydraulic cylinder chambers, 3) the effect of friction build-up on the moving table of the cutter. Condition monitoring of industrial machines Mr. Srinivas Raman M.A.Sc Candidate Supervisor: Dr. C.W. de Silva My research focuses on condition monitoring of industrial machines using wavelet packets and intelligent multisensor fusion. In industries where machine health is crucial to the production process and company revenue, it can be beneficial to implement a sophisticated condition monitoring system that is capable of prognosing and diagnosing faults in machines. If the machines are complex, numerous sensors maybe required to gather data and there needs to be a robust signal processing and classification scheme to accurately identify faulty conditions. I am current developing a system that uses wavelet packets to processes the acquired data, evolutionary computing to optimize feature vector selection and neural networks to classify the fault. My testbed is the fish cutting machine developed at the IAL, which represents a complex industrial machine with multiple functional units and subsystems. Machine Learning Based Multi-Robot Cooperative Transportation of Objects Gamini Dilupa Siriwardana M.A.Sc. Candidate Supervisor: Dr. Clarence W. de Silva This research project is a part of the main project of “Cooperative Multi Robots Systems for Emergency Response” and the aim of this project is to develop cooperative and intelligent control of autonomous multi-robot system in a dynamic, unstructured and unknown environment. Main objectives of my project is to study the problem of multi-robot object transportation, develop a physical multi robot system to cooperatively transport an assemble device to a goal location and evaluate the performance with respect to a benchmark engineering application available in our industrial automation laboratory. Machine learning technologies are employed to make decisions in the multi-robot system to cope with uncertainties in a dynamic and unknown environment. A reinforcement learning algorithm called the Q-learning, is used to solve the challenging problems of low learning speed and behavior conflicts in multi-robot learning. Local intelligence and mutual communication make system robust to erroneous perception or malfunctioning of mobile robots. C++ based simulation system and a physical multi-robot experimental system need to be developed to transport a device of interest to a goal location in a complex obstacle distribution environment. The developing approaches are implementing in the prototype system and rigorously will test and validate though computer simulation and experimentation Nano-electromechanical Vibration Sensor and Casimir Effect Dr. Farbod Khoshnoud Supervisor: Dr. Clarence W. de Silva Design and analysis of an embedded, nano-electromechanical capacitive sensor for vibration monitoring is under investigation in this research. In the sensor, vibration sensing is carried out by detecting the oscillations of a Single Walled Carbon Nanotube. Casimir effect due to quantum fluctuations in the zero point electromagnetic field is studied in designing the sensor. This device is particularly useful for precise and effective sensing of vibration for condition monitoring and fault diagnosis in machinery. Automation and Control in Manufacturing Technologies Dr. Ricky Min Fan Lee Supervisor: Dr. Clarence W. de Silva Research on product and process development in the Energy, Manufacturing and Resource sectors. Development on core and collaborative research projects with concentration on strategic technology areas with the highest potential impact on the above sectors in Canada. These include technologies to develop or improve manufacturing processes at the component level as well as at the machine and systems level that are crucial to the competitiveness of the Canadian industry. It is hoped to address the flexibility and rapid response time for the development of new products and processes, and reconfigurable manufacturing systems. The target research direction is as follows: Automation on manufacturing of fuel cells and other renewable energy component; Automation in EMS (Electronic Manufacturing Service), PCBA (Printed Circuit Board Assembly) and SI (System Integration). . |
