Intgrated manufacturing and production automation for the ceramic tile industry

The developed system is based on IR thermal imaging for the detection of temperature gradients due to internal delaminations. The thermal images are processed and classified by neural networks, in order to automatically have information about the integrity status of the tile (delaminated or not). This information should be used for the regulation of the press parameters. The system has been completed and successfully tested on-line. A diagnostic system based on high-resolution CCD cameras and image processing by neural networks was developed also for the early detection (after the drier) of broken corners and cracks. This system was successfully tested in the lab. The developed methodologies can be utilised also in other production technologies, as in the quality control for composite materials.

A three-channel telecentric flying spot scanner was designed and built to inspect embossed tile surfaces. The optical system contains a rotating multi-facet mirror and a large F#4 parabolic mirror of diameter 500mm. This scanner is capable of producing high-resolution (2048x2048 pixels) pseudo-colour images through a linear combination of scattered and specular light signals observed at three different angles. Ceramic tiles having different surface characteristics and various degrees of embossed features were scanned. Surface defects could be identified when the observation was made at non-specular angles where a colour camera fails to reveal these defects. Subtle variations of the surface profiles could only be seen with the multichannel scanner. The variation of the surface texture was easily identified in the pseudo-colour images. Finer details could be seen when the illumination was not normal to the surface. A 3-D Lambertian model was developed to obtain the profile height information, firstly in picture elements� slope to the detector and then in integrated slopes. The Laplace�s equation as in the second order elliptical PDE was implemented to check the surface integrity and also to evaluate unknown or inaccurate heights deduced from a specular or a noisy image.

The developed system is based on high-resolution cameras and image processing for grain size distribution measurement. The control and processing software was developed in LabView environment. The employed technique is a well-established one, but it was never applied for on-line applications on ceramic atomised powder. The system has been completed and successfully tested on-line. Also comparisons with the most expensive laser techniques have been performed and differences below 1% were found.

A colour and multispectral quality control measure for planar objects such as ceramic tiles, stone or organic material has been developed. The system uses optimal characteristics for the three input channels and expresses the data as three single dimensional arrays of 256 numbers. These numbers express the colours present on the surface and their relevant occurrences with a high signal to noise ratio. Additional proprietary hardware will enable the system to process the information fast enough to be used on-line. The system is to be used for analysis of very subtle colours. Up to now there have been very few ways to enhance the colours and contrast in a useful way. Histogram equalisation works on monochrome images but using the technique on each of the red, blue and green channels because of the three dimensional nature of the colour space. A new algorithm has been developed that enhances the colours and contrast in a controlled and predictable way. The technique although computationally intensive is suitable for production use with the right hardware. Both techniques are novel and have wide applications to fields outside the tile industry.

This is a reliable network capable of handling all types of process data, from activation of simple sensors to high-level information from computer based equipment such as the ATIS: Integrale, having the ability to connect to any fieldbus that may already exist in the factory. It incorporates an intelligent fieldbus card, which is capable of many fieldbus protocols, including Profibus, ModBus, Interbus-S and others. The acquired data is made accessible from any location within, or even outside, the factory. Java language is used for user interfaces so that it can run on any platform within the factory, and code can be executed in a web browser with no necessity for storing information or software on the computer that is accessing the monitoring system. This is particularly advantageous because it means that the monitoring system can be used from any computer within the network, without having to install software on each client computer individually.

The infrared sensor is currently used in the ceramic factory, but some limits have been found about its use, especially for on-line applications. For example, in the case of the porcelain stoneware production, some problems arise from the influence of the powder colour. In this research efforts have been spent in order to achieve a satisfactory calibration of the sensor and to standardise this procedure; moreover, it was demonstrated that it is possible to work with different calibration lines for each powder colour. The measuring system is actually installed at the end users plant. The developed methodology can be utilised also in other production technologies, as in the food or paper industry.

This system allows individual marking of the tiles at the exit of the press and then reading the identification back from the tile at relevant reference points. The whole system operates on-line. The information is used to cross-referencing tile quality-status to machine status, and to monitor production and identify parameters causing defective tiles. In addition the system provides the benefit that each individual tile can be traced back to the particular press, die and the time it was produced. The ID format used can provide enough combinations to mark each individual tile produced by a factory over any required period of time. Complete product traceability is possible and the information is available to MRP systems.

The moisture content in the spray-dried powder is one of the most critical output variables of the spray drying process. An infrared sensor has been tested and installed on line and the output measure is used to implement a control action on the spray drier operating temperature. Firstly a PI control was implemented and then the modular controller implemented by another partner has performed the control. A differential pressure cell transducer measures the pressure, and then a PI controller regulates the pressurisation baffle. A second closed loop control implemented via the modular controller on this process provides a constant value of depressurisation inside the drying unit. The two closed loop controls have significantly improved the quality of the spray-dried powder in terms of constancy of the most critical output variables.

The quality of the recycled water, which is introduced in the grinding process, is very important for the quality of the slip coming out from the mills. For this reason we have introduced some on line controls: a pH-meter, a conductivity-meter and a turbidity-meter. The first two measurements enabled the implementation of a closed loop control of the process, by addition of acid or base to correct the pH and dilution with pure water to correct conductivity, while the third instrument only gives an alarm if the value is higher than the set point. The control of the recycled water parameters has given the possibility to utilise it in the grinding process in addition to pure water. All the variables are collected and elaborated by a local supervision system.

This system allows automatic on-line tile classification and sorting of finished products. It is composed of two distinct modules, one module grades tiles in terms of their surface characteristics (tone and texture) and the other classifies tiles in terms of fault severity and frequency. The defects module is capable of detecting very shallow defects on highly reflecting surfaces that humans cannot detect at typical belt speeds and viewing angles. Class and grade of each tile are combined together, according to the specific production requirements, in one unique result and the system forwards this information to the tile-stacking machine that groups together all tiles of similar grade and class prior to packing. The modules operate at a belt speed of 1m/sec., on any texture and sizes between 100mm to 600mm width and any length. The modules are based on proprietary software/hardware architecture that allows concurrent image analysis operations. The systems are adaptively controlled to ensure stable operating environment and consistent image for analysis.

This is a scalable model for managing and accessing the data acquisition processes provided by the Monitoring System. All the data acquisition is run on Linux, except higher-level devices, which are directly connected into the system architecture that run on CORBA. The smallest object presented at the CORBA level is the machine. Each machine provides methods for access to its variables, and also basic identification. The object model allows 'machines' to be run on multiple computers throughout the factory, and to obtain the Object Reference required to locate the particular machine in which users are interested, or to list all the available machines in the network, for generic browsing. Measurements give production trends and are stored chronologically. When these are coupled with the tile tracking information the conditions that any given tile experienced as it passed through the machine can easily be ascertained. Correlation can be performed when the users require the information.

Tile tracking of unmarked tiles involves creating a model of the flow of tiles along the belt that allows the prediction of the position of all the tiles on the monitored section of belt. This is achieved with a number of proximity sensors, a shaft encoder and a micro-controller. The micro-controller can be trained so as to know the length of the section to be monitored and the size of the tiles passing along the belt. This information allows features of the tile flow to be monitored. It can monitor: - Production flow - Conveyor belt status - Tile status on belts It can control external devices for specific actions. It raises alarms on status malfunction. The system can be configured to monitor production flow across multiple belts. For multiple belts in inaccessible regions it is possible to set up a master/slave relationship between these units. The minimum system has two sensors expandable to eleven sensors enabling monitoring of very long belts.

The exhausted air coming from the firing and heating zone is aspired by a fan towards the pre-heating, and then it goes through a depurator that provides for the abatement of dust and fluorides. The regulation of the amount of aspirated air is very important for the firing process (tiles quality and depurator functioning): A closed loop control of the aspiration was implemented by the measurement of the depressurisation inside the kiln using a differential pressure transducer. The measured pressure is compared to the pressure set point and a PID controller regulates the aspirator fan by inverter. The set point is an important parameter of the firing receipt of each product. The advantages that we have found with respect to open loop process are: - Lower consumption of fuel; - Better operating of the scrubber; - Reduction of dimensional defects.

A Preventative Maintenance (PM) scheduler allows an integrated way of predicting, notifying and controlling PM events. An event is simply some task, which should be performed by factory staff in order to allow production at the highest level of efficiency possible. Events have a unique identifier, an associated date, a textual description, a severity level, a record of the machine to which they refer, and other status information. The scheduler is implemented within the Impact database. The architecture allows for monitoring alarms to be treated in the same way as other PM events. Typically these would carry a higher severity level, and may be notified to the end user differently, but the underlying database is used in all cases. In the Impact implementation, events can be defined either by entering a priori knowledge into the user interface, or by automatic notification of event from the monitoring system itself. Full automatic notification of PM events has been implemented in the ATISTM: Integrale product, and these events are passed into the Impact PM scheduler. It is envisaged that future work on a communications standard will incorporate an industry-standard way of describing PM events.

A system for the on-line density measurement of green tiles at the exit of the drier was developed and tested on-line. The system is based on a novel non-contact ultrasound measurement system (produced in the USA), used in this research to measure the propagation velocity and the attenuation within the tiles. These quantities are acquired and related to the density through a calibration line. It is the first time that a non-contact ultrasound system is applied on-line for quality control and monitoring. The on-line tests gave preliminary satisfactory results.

Livrables

Partager cette page

Télécharger