الفهرس 
Key Functions in Manufacturing ProcessOnly 14 pages are availabe for public view
 |  | from | 211 |  |  |
| | | from | 211 | | |
Abstract
Automation and integration of manufacturing systems has become essential to meet modern manufacturing requirements. Using new trends such as Artificial Neural Networks (ANNs) and Machine Vison (MV) have helped in enhancing manufacturing support systems (MSSs) for different production functions including design, planning, manufacturing and inspection. However, the large number of surveyed research did not address integrate CAPP and CAI tasks and their scopes were limited to rotational symmetrical parts with little research works on assembly parts inspection.
The aim of this research is to develop new system (CAI-X) for prediction of optimal machining conditions using ANNs and MV techniques. Part Images are used as main input for extracting key criteria and features of the examined part and then, using them in prediction of required cutting conditions needed for producing similar parts with similar criteria.
Digital Image Processing techniques (DIP) are used in image acquisition, enhancement, preprocessing and analysis to extract key data and attributes in part image. ANNs use these extracted data and attributes from part image in performing different tasks of MSSs such as CAPP, CAM and CAI systems. The ultimate goal is to develop an intelligent MSS able to predict and verify suitable machining conditions to produce the similar parts based on images of existing one.
The developed system consists of seven modules to perform the following tasks: material recognition, calculation of required raw material size and selection of suitable standard size; extraction and recognition of part features, evaluation of part surface roughness, process planning for selection of required machining conditions; and inspection of parts including surface and assembly inspections.
System performance was tested and verified by examining three different parts as case studies that examined the functions of different modules. Results indicated high accuracy and effectiveness of developed system in prediction and verification of optimal machining conditions for producing parts with given features and design criteria using image as main input. CAI-X has modular design which simplifies its use, modification while increasing provided options for users to use different modules either jointly or separately.
The advantages of the developed system are combining advantages of ANNs and MV while reducing their limitations. The main advantages include using non-contact methods based on images of inspected parts; in-process inspection applicability; flexibility and simplicity of use; ability to enhance performance and capture new knowledge. CAI-X is featured also with low human intervention in planning and inspection tasks, so it is less prone to human errors; saving costs and time consumed in part setups or fabrication of special fixtures needed in traditional systems and; suitability for mass production. The limitations of CAI-X are high dependency of results accuracy on image processing techniques and accordingly environmental factors like illumination system and machining noises. The developed system cannot be in inspection nor measurement of non-through internal features due to vision system limitations.