Integration of Taguchi’s method and multiple-input, multiple-output ANFIS inverse model for the optimal design of a water-cooled condenser

Significance Statement

Heat-conduction problems in high-heat-load applications remains a crucial aspect in design of a condenser. Although, experimental and analytical analysis has provided knowledge on heat conduction and flow fluid measurement, apparatus errors still remains a major challenge. Hence, constant improvement in model and test is necessary.

Taguchi’s method shows clear advantages over multiple-input, multiple-output systems in terms of appropriate assertion and implementation of products improvements, processes and equipment. Other factors such as temperature, humidity, fluid viscosity and major noise source could also be considered in Taguchi’s method during early stages of product design cycle.

Professor Chung-Neng Huang and and Dr. Cheng-Ching Yu from the Graduate Institute of Mechatronic System Engineering at National University of Taiwan presented an optimal design method for a water-cooled condenser which involved integration of Taguchi’s method and a multi-adaptive neuro-fuzzy inference system-based inverse method. The work published in journal, Applied Thermal Engineering considered a simulation analysis using COMSOL MultiPhysics software for verifying the effectiveness of the proposed method.

A neutral-network-based inverse model used in determining the relationship between outputs to manufacturing inputs despite its advantages does not converge well. However, an adaptive neuro-fuzzy inference system offers a combined advantage of neural system and fuzzy logic systems in effective modelling of non-linear systems.

The authors used shell-and-tube condenser in their studies. For platform development, one-fourth of original model was adopted as the simulation platform. A constant inlet flow speed and constant working-fluid flow temperature was set for initial conditions, convectional flux boundary was the outlet, the temperature boundary of hot runners was isothermal, stainless AISI 4340 was used as the solid heat conductor, contact surfaces of hot runners and fins were set as heat walls while parameters such as temperature of inlet (298K) and hot runner (325.85K), speed of inlet water (0.0625m/s), water density (988Kg/m3), sectional area of inlet and outlet (0.01252 m2) and diameter of hot runner (0.00635m) were considered.

Initial results shown by COMSOL MultiPhysics (original design) temperatures at specific points (p1-p8) showed average temperature of 303.98K and difference between highest temperature (p8, 305.12K) and lowest temperature (p1, 302.58K) was 2.73K. Comparing the above results with heat distribution values confirmed heat accumulation occurred. Difference between inlet and outlet temperatures is only 2.61K indicating low heat-exchange effectiveness.

For optimal design, Taguchi’s method was used to overcome large number of computations and high cost incurred by using full factorial method. Taguchi’s steps include cross validation in which an orthogonal array which provides sufficient information for training inverse model using minimum number of experiments was conducted followed by COMSOL-simulator-based experiment platform for determining heat distribution values or multiple-input multiple-output data pairs, manufacturing factors and temperature at specific points.

Data pairs and multi-adaptive neuro-fuzzy inference system were finally used to obtain a model describing the inverse relationship between the manufacturing factors and output temperature at specific points.

The simulation results showed that heat distribution had more uniform optimal design when compared to original which similarly indicates reduction in heat accumulation. Average temperature of points (p1 –p8) is 299.5K and difference between highest temperature (p1, 300.25K) and lowest temperature (p7, 298.4K) is 1.84K which is lower than the original design.

This study proved that water-cooled condenser based on optimal designs shows greater potential for heat exchange.

Integration of Taguchi's method and multiple-input, multiple-output ANFIS inverse model for the optimal design of a water-cooled condenser. Advances in Engineering

About the author

Prof. Chung-Neng Huang got Ph.D and master degrees in 2000 and 1997 respectively from Tohoku University, Sendai, Japan. Before that, he was a senior notional officer, working in Taipei city government for more than two years. Wherein, he had involved many key projects of Taipei city reconstructions. He got his bachelor degree in 1992 from National Taiwan University of Science and Technology. His current academic research, all undertaken in collaboration with industry, Optimal Design and Manufacturing, Renewable Energy, Hybrid System and Intelligent Control.

In recent years, he was invited to be the keynote speakers by the conference committees as; the 2nd/2013 International Conference on Applied Mechanics and Materials (ICAMM 2013) in Zhuhai, China, 23-24 Nov., 2013, The 2015 International Conference on Advances in Energy Resources and Environment Engineering (ICAESEE 2015) in Daqing City, Heilongjiang Province, China, 14-15 Mar., 2015, and the 2016 International Conference on Advances in Energy and Environment Research (ICAEER 2016), in Guangzhou, China. Furthermore, He was invited to be the committee, panel members and committee chairs.

Prof. Huang holds the external examiners to a number of leading academic and government departments in Taiwan and the memberships of the Institute of Electrical and Electronics Engineers (IEEE) and the Institute of Electrical Engineers of Japan (IEEJ). He has a long experience of acting as a consultant to industry in the fields of control technology, system automation, and new energy applications, etc.

Now, he is the editor-in-chief of the International Journal of Science and Engineering (IJSE). He owns over 20 patents and has published over 80 papers in academic journals and international conferences. 

Journal Reference

Chung-Neng Huang,, Cheng-Chung Yu.  Integration of Taguchi’s Method and Multiple-Input, Multiple-Output ANFIS Inverse Model for the Optimal Design of a Water-Cooled Condenser. Applied Thermal Engineering, Volume 98, 5 April 2016, Pages 605–609.

Graduate Institute of Mechatronic System Engineering, National University of Tainan, 33, Sec. 2, Shu-Lin St., Tainan 70005, Taiwan.

 

 

Go To Applied Thermal Engineering

 

 

 

Check Also

Fatigue Crack Growth Behavior of WAAM Steel Plates: Experimental Analysis and Comparative Study - Advances in Engineering

Fatigue Crack Growth Behavior of WAAM Steel Plates: Experimental Analysis and Comparative Study