Detailed defect information from a high-speed inspection system helps a coated groundwood producer eliminate costly breaks and bottlenecks Automated Inspection System Data Clarifies Runnability, Quality Issues Paper producers are expected to provide consistent quality from shipment to shipment, and inconsistent quality can create the risk of losing a customer. Whether producing paper in the mill or converting paper off site, the occurrence of material defects such as holes, coating voids, contamination, crush outs, and edge cracks is a major problem. Defects mean increased scrap, wasted energy, customer complaints, lower throughput, and less machine uptime. The costs of returns, rework and downtime are unacceptable in a highly competitive global marketplace. Over the years, paper mills have become familiar with the concept of basic defect detection. Many hole detection systems have been installed as a first step line toward implementing sheet inspection, and some mills have gone a step further by installing basic defect detection systems. These systems are at least able to provide some sort of real time information about the quality of the sheet while in production. The challenge remains to implement corrective process changes rather than simply monitoring quality. It is now possible to go beyond basic inspection for quality control and begin the integration of surface inspection systems into the operational side of the mill. By achieving a much higher degree of detection, resolution and classification, information generated by auto- Page 1
mated inspection systems represents the next level of performance necessary for today s competitive requirements. Many paper manufacturers have adopted automated inspection as a means of providing 100% real-time inspection. The integration of inspection technology has begun to extend from the paper machine and now includes coaters, supercalendars and secondary converting operations. Whether inspecting the base sheet, coating lines, slitting and sheeting operation or preparing the sheet for print, inspection systems have the potential to supply valuable savings by reducing breaks, rejecting defective sheets or saving expensive printing blanket damage. These factors alone are usually enough economic justification to support the purchase of the systems. The systems not only provide early warning of defects to the equipment operator, but also provide statistical information to more effectively manage the process and to increase production of a high-quality sheet. Today, automated web inspection systems must go beyond defect detection and provide accurate classification for root cause analysis, trend data for process improvement, roll data to optimize slitting, and a rewind control system for increased throughput. The following case study shows how one mill turned web inspection into a web improvement and cost reduction system. Coated paper producer aims for better quality For a major producer of coated groundwood, automated web inspection has offered a means to achieve better quality and runnability. The web width is 280 in., and the production line runs in excess of 3,800 fpm. The mill specified that the system had to detect holes and spots/crush outs of 0.5 mm² and edge cracks of 3.0 5.0 mm and larger, depending on edge stabilization. While the mill already had Page 2
experience with inspection systems on this production line, it was never able to get the necessary defect information to implement some sort of direct process control. The mill needed to identify eversmaller defects, but more importantly, it needed specific information about the different types of defects causing machine runnability problems and coater breaks. The new automated inspection system needed to provide alarms based on defect types and /or counts. It would also need to supply outputs linked to a marking system that could apply registration and defect markers. For process improvement purposes, the system had to classify defects by the combination of variable features including type, size, density, optical characteristics, time, and specific location. For this application, the mill selected ISRA SMASH system. This system runs at 60MHz, with each camera scanning at approximately 58,000 scans/sec. Until very recently, most inspection systems operated at 20 MHz or 40 MHz, with limited camera selection. Speed for the sake of speed has little value, but in an inspection system, speed translates into allocation of more pixels for inspecting the sheet at any given point in time. The greater the number of pixels on the sheet, the better the defect detection is, and more importantly, the better the image for classification analysis. The higher speed also allows the system to optimize performance based upon the surface area of the photosensitive pixel elements and the basic pixel geometry. In the case of this coated groundwood producer, a lower density pixel element was selected to increase the surface area for more sensitivity and retain a more square pixel image for true classification. This resulted in a very high pixel resolution for the defects and clear defect images for analysis and operator verification. Addition- Page 3
ally, this system included a formation analysis software module that provides a relative formation index of the sheet. The higher resolution images supply more detail so that formation variability could be reported by individual lane as well as the overall sheet formation. A stabilizing roll was installed on the retractable light beam to provide improved edge crack detection and allow greater access to the machine. The result was better detection on the edges and differing levels of classification based upon severity and trending to reduce tear offs. The special, variable high intensity LED light line provides illumination for high-speed production and can be optimized to achieve a specific light level for each grade. In combination with an ultra white light source, the light line provided additional classification data that had historically had obscured by the mills earlier off-white lighting. The light source has a uniform output and a long life. The combination of stability and long life makes the system very easy to access and maintain. From the paper machine to the coater The mill s new automated inspection system identifies defects, classifies them, and maps them to the hard drive, disk, or network. The edge-marking device then flags the defect for removal on the rereeler. The operating system setup proved easy, with menu-driven product selection and changeover without interrupting inspection. The inspection system also has individual defect visualization for immediate defect analysis. Whenever a defect image is displayed, it is accompanied by information on size, location and other classification criteria. The defect image and classification data are stored on the hard drive and along with roll reports, defect types, classes, size and Page 4
location, and can be transferred real time via Ethernet to remote locations for reel comparison. Early feedback from the system has enabled mill operators to make corrections before significant amounts off-quality paper are produced. Since the system accurately determines the types of defects within the reel, they can be optimally segregated for re-reeling. The finished reel is shipped to the re-reeler, where the operator utilizes the re-reeler control manager system to stop on the defect for removal or patching. Additionally, the re-reeler operator has the ability to preview selected defects to insure optimum throughput. These pre-selected stops are based upon product tables that are loaded into the system. The operator always has access to review both the pre-selected defects and any other defect included on the defect report. The combination of the accurate classification of defects with the inspection system and the re-reeler control manager system results in automatic control of the re-reeler for much improved mill efficiency. At this point, the coater operator can run with total confidence based on knowledge of what defects are remaining within the sheet. The operator knows what type of action has been taken at the re-reeler and can run the coater accordingly. This integration of inspection information from paper machine to re-reeler to coater allows the mill to manage the reel through the entire production process. In addition, real-time inspection information is available on the PCbased mill process control system with the RT Viewer software. From multiple sites within the mill, the system can be queried for both real time and historical data without any effect upon the inspection system and the operator s display. RT Viewer displays current Page 5
and historical reel data in the same format as the inspection user interface. Reel maps, trend displays, defect images, sheet opacity, formation statistics, and diagnostic information are also available at multiple locations. Historical inspection information is accessed from a secure, remote dataset using data mining software. The Data Mining software module provides an SQL database with optional interface to generate data queries that span date/time, grade, and shift ranges. Comparative reporting of trend and summary information is automated with scheduled e-mail delivery of pre-defined report types. For example, a shift report is automatically sent to the mill s production manager when a shift change occurs, while the quality control manager may be simultaneously receives an entirely different report formatted to meet his exact reporting needs. Eliminating Bottleneck, Reducing Breaks Mills are often asked for quality assurance reports based on ISO criteria, but ISO certification alone does not assure a superior product for customers. By getting total control of the production line, a consistent, first quality sheet is produced and customer satisfaction and machine uptime have increased. For the coated paper producer, two critical benefits were realized immediately. The inspection system was able to eliminate the choke point that exists between the paper machine and the coater. Without the inspection system in place, the paper machine had to slow down so that the coater could catch up after breaks caused by holes, edge cracks and other defects caused the sheet to snap off. The second benefit was the reduction in coater breaks which cost the mill anywhere from $5,000-$10,000 per break. A target of 30% Page 6
reduction in coater breaks was achieved within the first few weeks following the installation of the inspection systems. Without process feedback and 100% inspection, it is unlikely this could have been achieved. Images The automated inspection system runs at 60 MHz, with each camera scanning at approximately 58,000 scans / sec. (1.jpg) Page 7
The system identifies defects, classifies them, and maps them to the hard drive, disk, or network. (2.jpg) Data mining software automates time-consuming and costly manual data analysis. (3.jpg) Page 8