Scholarlink

The presented paper deals with the use of T-S (Takagi-Sugeno) fuzzy systems generated by ANFIS (Adaptive neuro-fuzzy inference system) to model and predict four key parameters of a gas turbine: discharge pressure (P2), high-pressure turbine speed (TNH), low-pressure turbine speed (TNL), and exhaust gas temperature (T7). These turbines are widely used across various industries, including hydrocarbons, aerospace, and power generation. Three partitioning methods (Grid Partition, Subtractive Clustering, Fuzzy C-Means (FCM)) were compared to identify the most suitable structure for this type of problem. The models were trained and tested on real-world data. Model performance was evaluated using metrics such as root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination (R2).

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

Optimization of ANFIS-TS Models Parameters of a Two-Shaft Gas Turbin: A Comparative Study

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Abdelouahab Chebbah,
  • Ahmed Hafaifa,
  • Mourad Bachene

摘要

The presented paper deals with the use of T-S (Takagi-Sugeno) fuzzy systems generated by ANFIS (Adaptive neuro-fuzzy inference system) to model and predict four key parameters of a gas turbine: discharge pressure (P2), high-pressure turbine speed (TNH), low-pressure turbine speed (TNL), and exhaust gas temperature (T7). These turbines are widely used across various industries, including hydrocarbons, aerospace, and power generation. Three partitioning methods (Grid Partition, Subtractive Clustering, Fuzzy C-Means (FCM)) were compared to identify the most suitable structure for this type of problem. The models were trained and tested on real-world data. Model performance was evaluated using metrics such as root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination (R2).