Estimation of the Solid Circulation Rate in Circulating Fluidized Bed System Using Adaptive Neuro-Fuzzy Algorithm

Abstract

Circulating Fluidized Bed gasifiers are widely used in industry to convert solid fuel into liquid fuel. The Artificial Neural Network and neuro-fuzzy algorithm have immense potential to improve the efficiency of the gasifier. The main focus of this article is to implement the Artificial Neural Network and Adaptive Neuro-Fuzzy Inference System modeling approach to estimate solid circulation rate at high pressure in the Circulating Fluidized Bed gasifier. The experimental data is obtained on a laboratory scale prototype in the Chemical Engineering laboratory at COMSATS University Islamabad. The Artificial Neural Network and Adaptive Neuro-Fuzzy Inference System use four input features—pressure, single mean diameter, total valve opening and riser dp—and one output feature mass flow rate with multiple neurons in the hidden layers to estimate the flow of solid particles in the riser. Both Artificial Neural Network and Adaptive Neuro-Fuzzy Inference System model worked on 217 data samples and output results are compared based on their Mean Square Error, Regression analysis, Mean Absolute Error and Mean Absolute Percentage Error. The experimental results show the effectiveness of Adaptive Neuro-Fuzzy Inference System (Mean Square Error is 0.0519 and Regression analysis R2=1.0000), as it outperformed Artificial Neural Network in terms of accuracy (Mean Square Error is 1.0677 and Regression analysis R2=0.9806).

An Intelligent Site Selection Model for Hydrogen Refueling Stations Based on Fuzzy Comprehensive Evaluation and Artificial Neural Network—A Case Study of Shanghai

“…There is often a highly nonlinear mapping relationship between its input and output, and it is generally difficult to write its expression, so it is called "black box" [28].The characteristic of ANN is that it can store information or knowledge distribution in a large number of neurons or the whole system. It has the potential of self-learning and self-organization [65]. In addition, it has strong fault tolerance and can deal with noisy or incomplete data [27,66].…”