RESEARCH ON PROCESS MONITORING AND CONTROL IN THE SEWING ROOM: A NOVEL APPROACH Helder Carvalho 1 , Ana Rocha 1 , João L. Monteiro 2 1 Dept of Textile Engineering, 2 Dept of Industrial Electronics University of Minho, Portugal email: hmc@eng.uminho.pt, amrocha@eng.uminho.pt, joao.monteiro@dei.uminho.pt Abstract: In the process of industrial sewing, set-up, tuning and monitoring still relies on empirical knowledge and subjective analysis of results. In order to respond to present market demands of small production runs and high quality products, real-time monitoring and control of the process are vital. Those techniques assure on-line detection of defects and their prediction, allowing set-up optimisation. To accomplish this objective, it is necessary to quantify the variables involved and to determine patterns of adequate and inadequate machine operation. This paper describes the development of a sewing test instrument, a valuable research tool for the development of new control and monitoring devices for sewing machines. This instrument is useful in “sewability testing”, in which textile materials can be tested for their ability to be sewed with quality. Copyright Controlo2000. Keywords: Set-up Optimisation, Real-time monitoring, Control, Textile Material Tests, Sewing Process Control. 1.INTRODUCTION In the apparel industry, sewing is one of the main processes, in which there has been a constant increase of the degree of automation. However, the process is not totally controlled. Mathematical models are normally unavailable and quantitative information about the operating parameters of the machines are in general unknown or not used in practice. Sewing process know-how is mainly empirical, based on apparel, sewing machine and accessories manufacturers’ experience; choice of needles, threads, and machine settings relies on general guidelines and trial-and-error. The increasing multiplicity of fabric structures, combined with the significant reduction of order size, boosts the need to reduce lead times and avoid quality problems, which normally introduce serious production delays. In this context, the apparel manufacturer can benefit from the use of equipment that can be quickly reconfigured and set up, or that (ideally) is self-regulated. Another aspect is the development of support systems to test the fabrics to evaluate machine set-up prior to production, as well as to help the textile manufacturer in guaranteeing fabrics “sewability”. Work in this area has started at University of Minho several years ago, with integration of sensors in the machines and signal acquisition solutions. The aim was to correlate those signals with the sewing process and finally extract relevant parameters of the sewing process (Ferreira, 1991; Rocha, 1996, Carvalho, M et al ., 1996a and 1996b). Later on, the work was focused on the improvement of automatic signal processing and result representation. Now, output of the system already displays results in a way that textile experts can decide on the quality of the sewing process, enabling the system to operate as a test rig (Carvalho, H. et al , 1997 and 1998b, Carvalho, H., 1998a). Other authors work on related fields, with close aims but different approaches (Zeto et al , 1995; Chmielowiec and Lloyd, 1995). The system has shown to be a valuable tool in the assessment of process behaviour, considering fabric properties, machine settings and configuration. The exact observation of the most important operating parameters is now possible, enabling the design of