A framework for automation of data recording, modelling, and optimal statistical control of production lines

• Ano de defesa: 2020
• Autor(a) principal: Leal, Flávio Murilo de Carvalho
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Federal do Cariri
• Palavras-chave em Português: Quality control; Automation; Statistical modelling; Embedded systems; Ciência da Computação
• Link de acesso: https://deposita.ibict.br/handle/deposita/420

Unarguably, the automation of data collection and subsequent statistical treatment enhance the quality of industrial management systems. The rise of accessible digital technologies has enabled the introduction of the Industry 4.0 pillars in Cariri local companies. Particularly, such practice positively contributes to the triple bottom line of sustainable development: People, Environment, and Economy. The present work aims to provide a general automated framework for data recording and statistical control of conveyor belts in production lines. The software has been developed in three layers: graphical user interface, in PHP language; database collection, search, and safeguard, in MySQL; computational statistics, in R; and hardware control, in C. The computational statistics are based on the combination of artificial neural nets and autoregressive integrated and moving average models, via minimal variance method. The hardware components are composed by open source hardware as Arduino based boards and modular or industrial sensors. Specifically, the embedded system is designed to constantly monitor and record a number of measurable characteristics of the conveyor belts (e.g. electric consumption and temperature), via a number of sensors, allowing both the computation of statistical control metrics and the evaluation of the quality of the production system. As a case study, the project makes use of a laminated limestone production line, located at the Mineral Technology Center, Nova Olinda, Ceará state, Brazil.