Many industrial plants use regulatory control loops to achieve a stable operation of its processes and ensure that their products meet minimum quality parameters required by their customers. Moreover, the transformation processes of the raw materials into finished products can involve complex organization of equipment and energy use of different types (for example, electrical, chemical, mechanical, thermal, etc.). In this context, a major effort is undertaken by companies so that your plants are operated efficiently. But for that to happen it is essential that regulatory control is working properly. The evaluation of the regulatory control can be made periodically by means of indicators, so that corrective actions can be taken when significant performance degradation thereof occurs. Considering these issues, it was decided to conduct a pilot project to analyze the benefits that the regulatory control evaluation could bring to a company of the steel industry. The pilot project was conducted in a steam power plant belonging to the company. The approach of the methodology employed, the main items checked and the results achieved in this project are part of the scope of this text.1
This article addresses the problem of detection of oscillations and root cause analysis for a thermoelectric power plant. The well-known method to detect oscillations based on autocorrelation function is used, which detects the signals with similar frequencies of oscillation and energy greater than a given threshold in this frequency band. Considering that these clustered signals are somehow related, an algorithm based on non-parametric identification has been proposed to indicate the root cause of this oscillation by finding the direction of the dependence between the signals. This approach has the advantage of not requiring the model order, dead time estimation and the definition of a model structure. The method is illustrated using a simulation example and also data collected from an industrial plant, with results validated using the process knowledge.Keywords Identificação de sistemas, Diagnóstico de falhas, Detecção de Oscilações, Dynamic Modelling, Time Series.Resumo Este artigo trata o problema de detecção de oscilações e análise da causa raiz de uma planta termoelétrica industrial. Para detectar as oscilações é usado o método já conhecido baseado na função de auto-correlação, que detecta os sinais com frequências de oscilação semelhantes, e que contenham energia maior que um certo limiar nesta banda de frequências. Considerando que estes grupos de sinais estão relacionados de alguma forma, um algoritmo baseado em identificação não paramétrica é proposto para indicar a causa raiz desta oscilação encontrando a direção da dependência entre os sinais. Esta abordagem tem a grande vantagem de dispensar estimativas da ordem do modelo, tempo morto e escolha da estrutura do modelo. O método é ilustrado através de sua aplicação a um exemplo de simulação e a dados obtidos de uma planta termoelétrica industrial, sendo validado a partir do conhecimento do processo.
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.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.