Scholarlink

<p>In order to assess the real performance of complicated real systems and create simpler models that can be applied to model-based control, failure detection, and industrial application optimization, measured data in conjunction with statistical approaches are being used more and more. In this work, an N-CARARMA (Nonlinear Controlled Autoregressive Autoregressive Moving Average) model was developed to predict the overhead temperature of a distillation column in a separation system. The input and output data collected under normal operating conditions were divided into two sets: The first set was used to estimate the model parameters using the Generalized Extended Stochastic Gradient (GESG) and Generalized Extended Least Squares (GELS) algorithms, while the second set was used for validation. To ensure the accuracy and reliability of the model, three key criteria were used to select the best structure: Akaike’s information criterion (AIC), root-mean-square error (RMSE), and Nash–Sutcliffe efficiency (NSE). The results, based on real process data, showed that the chosen model structure can effectively predict the changes in the overhead temperature of the column with high accuracy. Specifically, the N-CARARMA model (GELS) achieved an RMSE of 0.0212, AIC of − 0.4956, and an NSE of 0.9862, confirming its ability to capture the nonlinear dynamics of the process with high precision.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Identification of Overhead Temperature in Binary Distillation Column Using the N-CARARMA Model

  • 期刊论文

  • 发表日期：

    2026年4月12日

查看全文
  • Mohamed Zouai,
  • Lakhdar Aggoun,
  • Yahya Chetouani

摘要

In order to assess the real performance of complicated real systems and create simpler models that can be applied to model-based control, failure detection, and industrial application optimization, measured data in conjunction with statistical approaches are being used more and more. In this work, an N-CARARMA (Nonlinear Controlled Autoregressive Autoregressive Moving Average) model was developed to predict the overhead temperature of a distillation column in a separation system. The input and output data collected under normal operating conditions were divided into two sets: The first set was used to estimate the model parameters using the Generalized Extended Stochastic Gradient (GESG) and Generalized Extended Least Squares (GELS) algorithms, while the second set was used for validation. To ensure the accuracy and reliability of the model, three key criteria were used to select the best structure: Akaike’s information criterion (AIC), root-mean-square error (RMSE), and Nash–Sutcliffe efficiency (NSE). The results, based on real process data, showed that the chosen model structure can effectively predict the changes in the overhead temperature of the column with high accuracy. Specifically, the N-CARARMA model (GELS) achieved an RMSE of 0.0212, AIC of − 0.4956, and an NSE of 0.9862, confirming its ability to capture the nonlinear dynamics of the process with high precision.