Spline Dynamic Matrix Controller : uma nova formulação de controladores preditivos lineares
AUTOR(ES)
Matheus de Souza Escobar
FONTE
IBICT - Instituto Brasileiro de Informação em Ciência e Tecnologia
DATA DE PUBLICAÇÃO
2012
RESUMO
In the field of dynamic simulation, an understanding of the dynamic behavior of chemical processes is important from both process design and process system perspectives. Several methodologies have been developed so far, such as Step Response (SR), Transfer Function (TF) and Discrete State Space representation (DSS). Such approaches, however, have some disadvantages, regarding simulation speed, computational load and fixed sampling. From this premise, a new approach is proposed, where the input parametrization by piecewise cubic splines is combined with state space representation in the continuous domain, guaranteeing, from few data that gathers all the process dynamic information, simulation along a certain horizon. Such representation is called Spline Dynamic Matrix (SDM), inspired in the likewise representation of the Dynamic Matrix (DM) (Cutler and Ramaker 1980) proposed in the early 80¿s. The main idea resides in a multiplication of the input variation by a matrix that can embrace all the dynamic data, allowing a fast output calculation. By using this methodology, a new linear model predictive controller (MPC) is also proposed, Spline Dynamic Matrix Controller (SDMC), where the variable diminishment could lead to a faster and more robust resolution. So that such propositions can be validated, results are presented regarding dynamic process simulation and comparison of the performance achieved by the MPC proposed in this dissertation. In order to accomplish that, one MPC based on SR representation, for a SISO (Single Input Single Output) case study, and another one based on state space representation, for a MIMO (Multiple Input Multiple Output) system, were used, confirming that the potential and versatility of SDM technique can be truly perceived.
ASSUNTO(S)
controle preditivo controle de processos químicos simulação dinâmica
ACESSO AO ARTIGO
http://hdl.handle.net/10183/61057Documentos Relacionados
- Abordagem robusta para controladores preditivos
- Uma nova metodologia de jogos dinÃmicos lineares quadrÃticos
- Análise e desenvolvimento de controladores preditivos multivariáveis baseados em multi-modelos bilineares
- Controladores robustos para sistemas lineares com parametros incertos
- Controladores robustos aplicados em robôs bípedes