A carregar...
Publicação
Estudo e Implementação de um Sistema de Inspeção de Grades
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 29.42 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
A presente dissertação foi desenvolvida no âmbito da unidade curricular de TEDI
(Tese/Dissertação) do Mestrado em Engenharia Electrotécnica e de Computadores
(MEEC) - Especialização em Sistemas Autónomos.
O controlo de qualidade é um dos principais setores de qualquer indústria da
área da produção. Na Super Bock Group, S.A. (SBG), tal não é exceção, tendo
sido realizado um estágio curricular com o objetivo de implementar um sistema de
inspeção de grades, de maneira a assegurar a qualidade dos seus produtos, nomeadamente,
as grades de cerveja que devem estar completamente cheias antes de serem
exportadas. Além disso, este sistema deveria ainda ter a capacidade de inspecionar
diferentes tipos de grades com diferentes capacidades e estruturas organizacionais.
Assim, após alguma discussão e análise das possíveis soluções, optou-se pela
utilização do sensor de visão 3D O3D302. Porém, dado o custo associado, decidiu-se
a implementação de duas soluções:
• Solução com PLC: inspeção recorrendo a software comercial fornecido pela empresa
do sensor. Posterior integração com um Programmable Logic Controller
(PLC) para remoção das grades defeituosas da linha de produção, assim como
a apresentação de dados relevantes numa Human-Machine Interface (HMI).
• CrateSense: inspeção recorrendo às funcionalidades básicas do sensor O3D302,
recolhendo apenas as distâncias obtidas pelo mesmo a cada leitura realizada.
Posteriormente, aplicação de diferentes scripts para a avaliação das grades e
respetiva apresentação dos resultados.
Começando pela primeira solução, foram desenvolvidos diferentes programas de
maneira a obter um sistema o mais completo possível. Primeiramente foram configuradas
aplicações de completeness check para os diferentes tipos de grade a inspecionar
(Super Bock (0,33cl), Super Bock mini (0,20cl) e Carlsberg (0,25cl)). De seguida,
recorreu-se a uma ligação Ethernet (protocolo de comunicação Profinet) para a comunicação
com um PLC que, após a interpretação das leituras obtidas pelo sensor,
avalia a respetiva grade e ativa um dispositivo pusher caso seja necessária a remoção
da mesma. Por último, o PLC envia ainda informações estatísticas para uma HMI
para que os responsáveis pela linha de produção possam estar sempre atualizados
relativamanete ao estado da inspeção.
Com o software pronto, procedeu-se a colocação do sistema na terceira linha
de produção da SBG. De maneira a posicionar corretamente o sensor, foi desenvolvido
um suporte especificamente para a ocasião, tendo sido ainda montado todo o
hardware numa caixa, para maior segurança e organização do sistema.
Posto isto, o sistema de inspeção foi devidamente testado e validado, conferindolhe
uma elevada eficácia na deteção de grades com falta de garrafas, o que contribuiu
para a sua utilização regularmente.
Relativamente à segunda solução, esta consiste na aquisição de distâncias pelo
sensor O3D302, simulando um sensor Time-of-Flight (TOF) de baixo custo, a partir
de uma ligação Ethernet (protocolo Transmission Control Protocol/Internet Protocol
(TCP/IP)). Assim, a partir de diferentes scripts desenvolvidos, estes dados são
devidamente trabalhados e transformados em point clouds, permitindo a utilização
de técnicas de processamento de imagem na deteção de grades com falta de garrafas.
Por último, a solução deve avaliar as grades e defini-las como cheias ou underfilled
para que, posteriormente, o pusher pudesse retirar as grades defeituosas da linha.
This dissertation was developed as part of the Master’s Degree in Electrical and Computer Enginnering - Specialization in Aunomous Systems. Quality control is one of the key departments in any production industry. In SBG, this is no exception, and a curricular intershipt was carried out with the goal of implementing a crate inspection system to ensure the quality of their products, this is, beer crates that must be completelly filled before being exported. Additionally, this system should be able to inspect different types of beer crates with different cappacities and structural layouts. Therefore, after some discussion and analysis of possible solutions, the decision was made to use the 3D vision sensor O3D302. However, due to the cost, two solutions were considered and implemented: • Solution with PLC: inspection using the commercial software ifmVisionAssistant, provided by the sensor company. Subsequent integration with a PLC in order to remove underfilled crates from the production line and to present relevant data on an HMI. • CrateSense: Inspection using only basic functionalities of the O3D302 sensor, collecting the distances obtained in each reading. Subsequently, different scripts were applied to evaluate the crates and show the results. Starting with the first solution, different programs were developed to create the most complete system possible. Initially, completeness check applications were configured for each type of crate (Super Bock (0,33cl), Super Bock mini (0,20cl) and Carlbserg (0,25cl)). Then, an Ethernet connection (Profinet) was used to communicate with a PLC, which, after interpreting the data obtained by the sensor, the pusher will be activated if an underfilled crate was detected. Finally, the PLC also send statistical information to an HMI so that production line staff can stay updated on the inspection system status. With the software ready, the inspection system was installed on SBG’s third production line. To position the sensor correctly, a custom support was developed, and all hardware was mounted in a box for safety and system organization. Following this, the inspection system was properly tested and validated, demonstrating high effectiveness in detecting crates with missing bottles, which contributed to this system’s regular use. As for the second solution, it involves acquiring distances from the O3D302 sensor, simulating a low-cost TOF sensor, through an Ethernet connection (TCP/IP protocol). Data from the sensor is processed and transformed into point clouds using different scripts, enabling the use of image processing techniques to detect crates with missing bottles. Lastly, the solution must evaluate the crates and classify them as full or underfilled so that the pusher can subsequently remove defective crates from the production line.
This dissertation was developed as part of the Master’s Degree in Electrical and Computer Enginnering - Specialization in Aunomous Systems. Quality control is one of the key departments in any production industry. In SBG, this is no exception, and a curricular intershipt was carried out with the goal of implementing a crate inspection system to ensure the quality of their products, this is, beer crates that must be completelly filled before being exported. Additionally, this system should be able to inspect different types of beer crates with different cappacities and structural layouts. Therefore, after some discussion and analysis of possible solutions, the decision was made to use the 3D vision sensor O3D302. However, due to the cost, two solutions were considered and implemented: • Solution with PLC: inspection using the commercial software ifmVisionAssistant, provided by the sensor company. Subsequent integration with a PLC in order to remove underfilled crates from the production line and to present relevant data on an HMI. • CrateSense: Inspection using only basic functionalities of the O3D302 sensor, collecting the distances obtained in each reading. Subsequently, different scripts were applied to evaluate the crates and show the results. Starting with the first solution, different programs were developed to create the most complete system possible. Initially, completeness check applications were configured for each type of crate (Super Bock (0,33cl), Super Bock mini (0,20cl) and Carlbserg (0,25cl)). Then, an Ethernet connection (Profinet) was used to communicate with a PLC, which, after interpreting the data obtained by the sensor, the pusher will be activated if an underfilled crate was detected. Finally, the PLC also send statistical information to an HMI so that production line staff can stay updated on the inspection system status. With the software ready, the inspection system was installed on SBG’s third production line. To position the sensor correctly, a custom support was developed, and all hardware was mounted in a box for safety and system organization. Following this, the inspection system was properly tested and validated, demonstrating high effectiveness in detecting crates with missing bottles, which contributed to this system’s regular use. As for the second solution, it involves acquiring distances from the O3D302 sensor, simulating a low-cost TOF sensor, through an Ethernet connection (TCP/IP protocol). Data from the sensor is processed and transformed into point clouds using different scripts, enabling the use of image processing techniques to detect crates with missing bottles. Lastly, the solution must evaluate the crates and classify them as full or underfilled so that the pusher can subsequently remove defective crates from the production line.
Descrição
Palavras-chave
crate inspection completeness check HMI ifmVisionAssistant O3D302 sensor PLC Python quality control Inspeção de grades Controlo de qualidade Sensor O3D302