Full project title: “Definition, Development, and Validation of Industry 4.0 Data Analytics Reference Architecture Model for Batch Processes“
Motivation
The successful and uninterrupted operation of textile dyeing process & machinery, fully automatic powder dyestuff weighing & distribution systems is one of the most important success criteria of textile factories in terms of cost, sustainability and competitiveness. Batch40 is poised to demonstrate direct and generalisable benefits for the project partners, namely Enforma & Eliar.
Project Objectives
Textile is one of the main sectors in the world and in Turkey. There are different stages in the textile production process. One particular process that consumes a lot of water, energy, chemicals, and dye is the dyeing stage. The basic principle in the dyeing stage is to achieve right-first-time.
In this particular industry where the traditional technologies and approaches are still widely used, fully automatic control of dyeing processes integrated with production system management leading to minimal human intervention & error is an important challenge. As well as validating a possible solution in the end user, generalising it to other areas of application is another goal of the project.
Thus, the aims of our project contains the following:
- Researching use case scenarios that will be the basis for analytical studies.
- Creation of the reference architecture model (RAM).
- Developing as a platform and field trials of ML/AI/MCDSS/visualisation & key FMECA components within the architecture.
- Ensuring the persistence of generalisation & project gains through feeding the results back to RAM.
Keywords
Advanced predictive analytics, multi-criteria decision support systems (MCDSS), manufacturing 4.0
Innovative aspects
Traditionally, textile dyeing factories have weak systematics for data analytics and solutions implemented are relatively primitive. Innovative aspects of Batch40 are as follows:
- Detailed analysis of textile dyeing process and dyeing machine situations, dye weighing & distribution processes depending on inputs and outputs.
- AI-based detection of anomalies & frequencies for multivariate operations, each with almost unique patterns.
- Development of reference architecture for R&D systematics with potential portability to different verticals.
- Connecting advanced predictive analytics studies with embedded systems & production management systems.
- Adaptation of the solution as edge, on-premises & cloud implementation.
Economic and national gains
National gains:
- Training of knowledgeable and experienced engineers in data analytics, IIoT & predictive maintenance technologies. Ensuring qualified employment increase in the manufacturing sector and the sectors affected by the multiplier effect.
- The potential of the products & solutions to be developed within the scope of the project to prevent imports in the field of IoT & manufacturing and to increase the export of high value-added technology.
- Potential to launch new projects in similar areas.
- Contributing to the scientific and technological development of our country by going the way of protecting IPR & making scientific publications on the subjects to be worked on in the project.
- The opportunity to expand abroad through the channel of deployment of the outcomes of the project in international markets.
Economic gains:
- In a factory with 30 dyeing machines, the techniques can potentially save ₺1.6M+ annually.
- There will be product and service income arising from the commercialisation of the platform.
Interim results
An interdisciplinary approach covering both the process control and the data science aspects of a textiles finishing process shows promising improvements. The outcomes are better optimised processes, reduction in energy & material & water consumptions, fewer undesired oscillations, and a higher rate of right-first-times.
