NONLINEAR GLOBAL MODELLING FOR DRUM-BOILER SYSTEM IN SUBCRITICAL COAL-FIRED POWER PLANT USING ADAPTIVE NEURO-FUZZY INFERENCE SYSTEM (ANFIS)
Drum-Boiler is one of main components in coal-fired power plant that used to produce steam. Produced steam then used to rotate the turbine that being coupled to the generator. As a result, electricity is generated. During operation, water level in drum vessel should be stably maintained as power dem...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/21433 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Drum-Boiler is one of main components in coal-fired power plant that used to produce steam. Produced steam then used to rotate the turbine that being coupled to the generator. As a result, electricity is generated. During operation, water level in drum vessel should be stably maintained as power demand change in order to prevent coal-fired power plant from being trip. Therefore, a model needed for describing the behaviour of drum-boiler system to maintain water level. <br />
<br />
<br />
<br />
<br />
In this research a model for drum-boiler system is developed using Adaptive Neuro-Fuzzy Inference System (ANFIS). This ANFIS model is developed with gey-box modelling approach. In this case, we are using some knowledge about pysical variable relationship that governed from physical conservation law. A mathematical equation that is Nonlinear AutoRegressive Exogenous (NARX) used on ANFIS’s artificial neural network so that global behaviour of the system can be describe. <br />
<br />
<br />
<br />
<br />
ANFIS model trained by using daily measurement data from subcritical coal-fired power plant. Model validation is done by using training data and testing data. Furthermore, model validation also done on two zones which is operating point zone and transition zone. Model Validation shows that developed model successfully describe system behaviour globally on several operating point zones and transition zones. In transition zone, model succesfully captures shrink and swell phenomena that occur on the system. Validation result on operating point zone gives RSME value of 3.8853 using training data and 3.8220 using testing data. While on transition zone, validation gives RSME value of 4.8878 for training data and 6.7319 for testing data. |
|---|