Implementasi Kontrol Model Prediksi Berbasis ANFIS Pada Mesin Penghasil Uap Air

Abstract

Steam generator has problem such as dependent variable on the output parameters, therefore it is needed a way to get system behavior of steam generator which is represented by model of the plant. Implementation of the model predictive control in this research was represented by using simulation, so the object that was controlled was virtual steam generator. In this research was used MATLAB software as a virtual steam generator and also for computes ANFIS, whereas the LabVIEW software was used as a representation of control room.

From the research, it was found the best parameters for each ANFIS that was used as a model unit in the model predictive control, that was by using historical data 4^th as much as 800 datas, the percentage ratio of learning for training data and checking data on each ANFIS for each model structures sequentially by 90% and 10%, except the percentage ratio for ANFIS on water flow parameter sequentially by 80% and 20%. The results of validation RMSE (Root Mean Square Error) by testing for 100 datas, it was obtained values as follows: water flow=1.9941, water pressure=48.0236, air flow=604.0621, fuel gas pressure=0.7087, fuel gas temperature=18.6594, O₂ content=0.9591, steam pressure=76.1557, steam quality=3.9734 and steam flow=264.9173.