The present work proposes a neural network model capable of anticipating possible faults in a semiconductor manufacturing plant by predicting non-linearity spikes in sensor data. Early detection of significant variation can be crucial for identifying machinery degradation or issues in the process itself. We use non-linearity as it is not affected by regular process changes and autocorrelation, thus avoiding false-positives in the neural network caused by changes in demand and the presence of control systems. The developed model is able to predict up to 30min of future non-linearity with loss ≤ 0.5. Furthermore, the proposed model is flexible enough to present itself as a starting point for future work in the field of fault detection in other areas.
O presente trabalho apresenta a aplicação de uma solução de controle avançado para sistemas HVAC dedicado ao condicionamento de ar de uma sala limpa do instituto itt- Chip/Unisinos. Sistemas HVAC (Heating Ventilation and Air Conditioning) são responsáveis por grande parcela dos custos de produção de semicondutores por conta da necessidade do controle de temperatura, umidade e pressão interna dos ambientes controlados. Por conta do grande potencial de contribuição ao mercado de semicondutores, foi desenvolvido um sistema de controle preditivo multivariável em nível de aplicação visando o atendimento das especificações de controle requeridas pelas salas limpas do instituto tecnológico. De modo a obter-se a síntese do controlador preditivo, foi realizada a caracterização dos processos psicrométricos e modelagem paramétrica a partir de ensaios de resposta. Foram desenvolvidos ensaios do comportamento servo e regulatório do controlador preditivo e a avaliação de um indicador de consumo energético (ECI) dos benefícios energéticos quando comparado com o controle PID originalmente instalado. Os resultados revelaram um bom comportamento servo e regulatório do controle MPC e uma redução no indicador de consumo energético que pode variar entre 80 e 90%, a depender da sintonia escolhida para o controle preditivo.
Cloud, IoT, big data, and artificial intelligence are currently very present in the industrial and academic areas, being drivers of technological revolution. Such concepts are closely related to Industry 4.0, which can be defined as the idea of a flexible, technological, and connected factory, encompassing the shop floor itself and its relationship between workers, the chain of supply, and final products. Some studies have already been developed to quantify a company’s level of maturity within the scope of Industry 4.0. However, there is a lack of a global and unique index that, by receiving as input how many implemented technologies a company has, enables its classification and therefore, comparison with other companies of the same genre. Thus, we present the I4.0I (Industry 4.0 Index), an index that allows companies to measure how far they are in Industry 4.0, enabling competitiveness between factories and stimulating economic and technological growth. To assess the method, companies in the technology sector received and answered a questionnaire in which they marked the technologies they used over the years and the income obtained. The results were used to compare the I4.0I with the profit measured in the same period, proving that the greater the use of technology, the greater the benefits for the company.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.