OPTIMIZATION OF ARTIFICIAL NEURAL NETWORK MODEL FOR IMPROVEMENT OF ARTIFICIAL INTELLIGENCE OF MANUALLY DRIVEN BRICK MAKING MACHINE POWERED BY HPFM
OPTIMIZATION OF ARTIFICIAL NEURAL NETWORK MODEL FOR IMPROVEMENT OF ARTIFICIAL INTELLIGENCE OF MANUALLY DRIVEN BRICK MAKING MACHINE POWERED BY HPFM
P. A. Chandak1 and J. P. Modak2 1Department of Mechanical Engineering, DMIETR, Wardha (MH), India 2 Professor Emeritus, Dean (R&D), PCE, Nagpur(MH), India
ABSTRACT
Considerable development
has been done by some authors of this paper towards development of manufacturing
process units energized by Human Powered Flywheel Motor (HPFM) as an energy
source. This machine system comprises three sub systems namely (i) HPFM (ii)
Torsionally Flexible Clutch (TFC) (iii) A Process Unit. Process unit so far
tried are mostly rural based such as brick making machine (both rectangular and
keyed cross sectioned), Low head water lifting, Wood turning, Wood strips
cutting, etc HPFM comprises pedalling system similar to bicycle, a speed rising
gear pair and a flywheel big enough such that a young lad of 21-25 years, 165cm
height, slim structure can pump energy around 30,000 N-m in minutes time. Once
such an energy is stored peddling is stopped and a special type of TFC is
engaged which very efficiently brings about momentum and energy transfer from
flywheel to a process unit. Process unit utilization time upon clutch
engagement is 5 to 15 seconds depending on the application. In other word
process units needing power of order of 3 hp to 10 hp can be powered by such a
machine concept. Experimental data base model is formulated for Human Powered
Flywheel Motor Energized Brick Making Machine (HPFMEBMM). The focus of the
present paper is on development of an optimum Artificial Neural Network (ANN)
model which will predict the experimental evidences accurately and precisely.
The optimization is acknowledged though variation of various parameters of ANN
topology like training algorithm, learning algorithm, size of hidden layer,
number of hidden layers, etc while training the network and accepting the best
value of that parameter. The paper also discusses the effects and results of
variation of various parameters on prediction of network.
KEYWORDS
ANN, Matlab,
Manually driven brick making machine. Simulation
Comments
Post a Comment