Exploitable results Roughing of leather using a blasting technique The proposed novel approach to roughing, i.e. the use of grit blasting techniques to remove the upper layers of the leather, requires a large capital investment to implement and this would clearly be a controlling factor on it's uptake in the sector. The quality monitoring technology developed within the project however is to be designed to be standalone and be applicable to any roughing system. This feature is seen as crucial as such a system could be used by any shoe maker and would potentially have a large market, and once implemented would reduce the costs involved in quality monitoring when implementing automated systems. The effectiveness of air propelled abrasives in removing the surface layers of the leather in order to optimise bond strengths were unknown at the time the project started. Whilst the grit blasting process will remove leather the resulting surface finish and the amount removed are crucial factors in the resulting bond strength. In addition it is essential that the process can operate at linear speeds appropriate to the speeds required for roughing a complete shoe, i.e., one shoe in 10 seconds. Preliminary trials using a simple linear test rig were arranged such that the various parameters (differing blast angles and grit speeds/ densities) involved could be varied in a controlled way and to enable samples of materials to be experimentally roughed. The process very soon proved that it was very viable. Further trials were carried out with both leather and synthetic samples. The surfaces were roughed with varying parameters and independently tested by each of the RTDs following the same test procedures. The samples were also examined using the proposed control sensors in order to determine a method of measurement, which would rank the samples in the same order of bond strength as were the results of the testing. Workpiece handling and positioning system The various sub systems developed in the IRIST Project must be able to come together and fit neatly and efficiently into the production environment. This joining could be made through the handling system, which should be include conveyor and shoe support to the work piece loading, handling and transportation. LIREL has a great diversity of conveyor systems. The equipment chosen was the one that presented the characteristics more adequate to this application. This system consists in a conveyor with a robot controlled by specific software developed by LIREL. The conveyor has two lines: the main line (outer line) and the secondary line (inner line). The inner line works as a store of the lasted uppers that were waiting for the treatment in the shot blasting. The robot will be responsible for the manipulation of the lasts from the conveyor to the shot blasting equipment. The two components, shot blasting chamber and robot, must be connected by a software to be possible the transmission of information between them. The software that already exists allows the following actions: - Configuration of the entry and exit of the lasted uppers in the conveyor, - Configuration of the relative positions of all components (shot blasting, robot, conveyor, shoe support, etc.) Integrated roughing system The surface treatment system developed in the IRIST Project use air-propelled abrasives (shotblasting) for the surface treatment of shoe uppers. The process conditions such as shot characteristics, velocities and angles of incidence are being studied and will be established. Obscuration sensor In order to provide control of the blasting system all the process variable effects on blasted leather surface quality must be quantified and suitable process feedback/control mechanisms identified. While pressure control of the air delivery and mass-flow control of the grit delivery processes is possible, using pneumatically operated gate valves, a consistent measure of the mass-flow rate at the blast head has been difficult to obtain. The development under this programme of an optical measurement system which can be easily calibrated to give a good indication of on-line grit mass flow rates has helped to alleviate this measurement task. The various system components are mounted in a purpose made enclosure that is attached to the end of the blast nozzle. The sensor system has been tested at the grit blasting facility at the University of Newcastle. A series of test were conducted to investigate the performance of the sensor with air only delivery and with various settings of the manual delivery valve. The sensor was able to distinguish between all five settings of the manual value for a range of air pressures. In addition the sensor was also able to determine sudden instantaneous drops in the grit delivery mass flow rate due to inconsistent delivery at the gate valve. The sensor system has been developed to a standalone version. Customized roughing tool path automation and handling system In order to control the process effectively through feedback from the sensors there has to be accurate control of the robot manipulator trajectory relative to the lasted upper and the shot blasting nozzle. The nozzle venturi is mounted on the end of a 40mm diameter high-pressure hose, which allows the nozzle to be positioned anywhere within the blast chamber. It was evident that the motion of the hose and the nozzle needs to be minimized to both reduce wear and tear on the hose and to reduce variability in the grit stream caused by excessive flexing of the hose. Each lasted upper has to be roughed in approximately 10 seconds and manipulating only the nozzle relative to the upper requires very rapid orientation of the robot end-effector, particularly around the toe and heel areas. Furthermore, the wrist articulation of the nozzle twists the hose, requiring the robot to effectively unwind itself at the end of each process. For this reason it was decided to dynamically manipulate the lasted upper and the nozzle, simultaneously. By rotating the lasted upper much reduced wrist orientations were required of the robot and greater control of the nozzle was realized, improving the process and longevity of the moving parts, as well as simplifying the robot trajectory. The laboratory demonstrator has been successfully used to grit blast a large number of shoe samples with a wide variety of style. Synchronizing the rotary table with the manipulator motion has brought about improved performance. The design of the handling system involved considerable evaluation. It was recognised that it had to capable of providing adequate support during both scanning operations and processing. The handling unit is mounted onto a servo-controlled two axis integrated rotary table and slide-way. The rotary axis is required to provide dynamic orientation of the work-piece during processing to avoid the need for the robot to have to undertake excessive rotations, which would otherwise be problematic for the unwieldy high-pressure hose on which the nozzle is attached. The linear slide is used solely for locating the work-piece in one of three positions, during operator mounting, scanning and grit blasting. A pneumatically operated clamping unit has been designed and constructed. The clamping unit is capable of accommodating a wide variety of shoe sizes and types. The localisation of the work-piece and integrated handling system is robust, and maintains its position accurately throughout the processing cycle. This feature is particularly important to ensure correct registration between the scanning operation and subsequent processing sequence. Thermographic based quality inspection system The system developed, is a quality cont l for roughed leather based on thermography sensors. This technology is protected by a patent. The main result is an industrial system that works by heating the surface of the leather and a thermographic sensor reads the surface temperature, and by using an informatics system we can control the quality of the roughed area by the comparison of the temperatures. Masking system During the early preliminary trials it was observed that the blasting process produced a distinctive pattern across the leather. The centre portion was well roughed and is bordered by a well-defined area, which forms a clear interface between the well-roughed area and the untouched leather. Depending on the angle of approach, there is a varying amount of damage to the surrounding grain reinforcing the need for varying degrees of protection to the upper material. A simple mask provided an extremely well defined edge to the roughed area and in fact may introduce a potential weakness into the leather. Experiments with non-contacting masks revealed that significant damage to the upper material did take place. Further experiments showed clearly that there needs to be actual contact between any sort of masking device and the upper material to produce the best roughing results. Even allowing small amounts of dust and debris to pass over the upper material results in unacceptable damage, typically resulting in a matting of the surface grain. Several different designs of mask were evaluated. The adopted solution used a rotating satellite disk of 60mm diameter, which traces a path around the feather line of the lasted upper both acting as a guide for the venture nozzle and as mask to protect the upper from over roughing. The disk is maintained in contact with the work-piece using passive compliance. Extensive tests have been carried out and the results obtained are considered excellent, with a well-defined transition between the processed area and the cosmetic surface showing no visible sign of abrasion. Scanning system and conformal mapping A structured light vision system has been designed and integrated with the IRIST demonstrator, in collaboration with RT&M. The lasted upper is scanned at variable intervals along its length and the scanned data is then processed to find the "best line" for each scan. Appropriate filtering is applied in order to smooth each line and the edges identified based on slope data. The filtering comprises smoothing and removal of any outliers or artefacts and then low pass/FIR filtering of the resultant line. This software was developed in order to provide a method of identifying the extent of the feather line on the roughed upper. This is accomplished by analysing the rate change in the slope data provided by the software over the region of interest. The set of edge pixels {(x,y,z)} from the boundary of the roughed area together with information on the normals to the surface are stored in an ASCII file. The surface normals are required to permit the blast nozzle to approach the {(x,y,z)} co-ordinate at a suitable angle and hence blast at the correct orientation to the lasted upper sole. The ASCII file is subsequently matched with a similar 'sole' map to obtain an optimum fit. The fitted coordinates are then processed using an interpolation algorithm and then converted to positional data from the scanner's plane to a reference frame, to provide suitable {(x,y,z)} and normal data for the robot used to control the blasting head. This will enable the blast nozzle to not only follow the correct tool path but also allow the blasting angle relative to the surface to be controlled. Concepts for matching the roughed area of the lasted upper to individual soles for flat, rounded and side wall-designed shoes have also been developed. This technique involves the following: 1. Virtually placing a scanned version of the sole onto the scanned upper in approximately the correct position and orientation. 2. Obtain measures of fit: These will include 'balance' measures and goodness of match of the curvatures in the blasting area. 3. Virtually adjust the placement to get a match to an acceptable level. In order to do this in a reasonable time the scanned data point set will have to be reduced to a minimum, concentrating on the areas where 'fit' is critical. The technique has been satisfactorily tested on a wide range of flat-soled shoes. Searching for OpenAIRE data... There was an error trying to search data from OpenAIRE No results available
