Smart Production

Quality step-up in yarns manufacturing by predictive models

In current years, the textile industry required more complex yarns, which are more difficult to produce. This creates challenges, because especially in the textile industry, process control is strongly based on expert knowledge and experience – the latter element does not exist for new production processes. Our Italian partner Pecci Filati is producing such complex yarn and has decided to compensate missing experience in new processes with a data-driven approach. Based on the DatenBerg technology the company will implement data-driven support for the operators on the shopfloor. This will enable the workers in finding the real root-cause of quality issues based on data and help them to identify scrap and quality deviations early on with prediction models. The pilot for this new approach is supported by the European light industries innovation and technology (ELIIT) project.

Problem description

Just like in all manufacturing companies’ quality errors occur in the textile production. One driver is the increasing customer demand in terms of complexity over the years. More complex products lead to more complex machines which are harder to control by operators with traditional methods (Experience). Right now, the quality of the products is heavily dependent on expert knowledge in the manufacturing site. Furthermore, all deviations on the quality are identified after manufacturing or even worse on the customer site. This means, actions against deviations always happen in the retro perspective. Pecci, with its high-quality products in a competitive market, is in big need of delivering stable quality and always improve its production. The two main challenges in textile companies can be summarized as following:

Dependency on expert knowledge and reactive acting: Quality is highly dependent on subjective expert knowledge, which today is crucial for the process steering. Actions against quality issues take place after the deviations have occurred. This leads to high monetary loss (scrap) and the threat of a negative perceived image of the manufacturing brand at the customer site.

Cause-and-effect: Tracing back quality failures to specific process parameters is quite difficult, even though newer machines provide increasingly more data. But, due to missing knowledge in data-analytics and lack of internal capacity in the IT department, a profound analysis cannot take place.

Solution approach

With new technology emerging, more data gets available in the manufacturing site. After several big innovation waves such as the lean manufacturing movement or Six Sigma, the next big thing is the so-called „Industry 4.0” (I4.0). While many different definitions of I4.0 do exist, a common denominator is among most of them data-driven improvements. Which is what the DatenBerg stands for. The data of new machineries will be used as a lighthouse project to show the potential of such analytical approaches for the textile world. With the pilot two use-cases will be implemented:

1) Root-cause identification: Support the operator by automatically finding the root causes of quality deviations. Anomalies and patterns leading to bad quality will be identified and presented to the operator in a easy understandable way.

2) Proactive actions: Predicting the quality of the produced part in near real-time based on process parameters will help to idenfity scrap and deviations early on. The operator will be enabled with a traffic light system to understand the current condition of the process.

Implementing this use-cases will drive the achieved quality level as well as support the operators on the line with data-based insights. During the next twelve months DatenBerg will implement these use-cases in Pecci’s manufacturing site located in Prato, Italy. Further support during the implementation will be provided by the Next Technology Tecnotessile (NTT) as well as from the ELIIT coordinator AITEX from Spain. During the project, several posts on the implementation will be published – so stay tuned for further information.

We are looking forward to support Pecci on their way towards a smart factory – A warm welcome in the DatenBerg team!

Involved partners

Pecci is a unique yarn producer based in Prato, Italy established in 2002. All yarns are produced internally in the production plant. Many different and modern technologies allow Pecci to produce all types of fancy yarns for knitwear and for handknitting, used by most important brands in the apparel world. Pecci produced over 500 000 kg of yarn last year with a headcount of 55 people. Pecci will act as the application partner, provide the necessary knowledge about production processes and its manufacturing site as a demonstrator.
(https://www.pecci1884.it/pecci-filati/)

Next Technology Tecnotessile (NTT) is a research and services organization, recognized by the Italian Ministry of Education, University and Research (MIUR), established in 1972 and based in Prato. As subcontracting activity, it will provide consulting to Pecci in managing the information which will be provided by DatenBerg after data analysis. Thanks to the long-time experience in the textile sector and its focus on Industry 4.0 innovation, NTT will support Pecci in the exploitation at shop floor level of DatenBerg analysis results.
(http://www.tecnotex.it/en/index.html)

European light industries innovation and technology project (ELIIT project) seeks to support textile, clothing, leather and footwear (TCLF) SMEs in enhancing their competitiveness while helping them integrate new technologies in innovative or high added-value products, processes or services. ELIIT will aid the transfer of innovation and technology as well as the market uptake of innovative solutions. We do so by developing concrete pilot actions to improve productivity, value chain integration, and resource efficiency. The described project got selected out of over 93 proposed partnerships.
(https://ec.europa.eu/growth/tools-databases/eliit_en)

Schreibe einen Kommentar

Deine E-Mail-Adresse wird nicht veröffentlicht. Erforderliche Felder sind mit * markiert.