Experimental and smart (ANN)‎ study for transient heat transfer from a flat plate under air impingment

الملخص EN

An experimental study is conducted to verify the transient heat transfer characteristic for a target plate under air impingement jet.

This investigation has been carried out with different sizes of orifice diameters included (D=5, 10, 15 and 20mm).

Air jet velocity is varied using a control valve and piping system designed for this purpose .

Air leaves the orifice and impinges the constant flux heated plate orthogonally .

Reynolds number is varied in the range (Re=7100-11900) representing air jet velocities in the range (Uj=18 ــ 40m/s).

The orifice to plate distance ratio is in the range (H/D=2, 4, 6 and 8).

The transient heat transfer coefficient is calculated using the semi-infinite solid equation.

The local heat transfer coefficients and local variation of Nusselt number are measured for the transient and steady state case, represented by the time period which represent the a dimensionless parameter for the the time elapsed from start of cooling process ( time =0and τ_o=0) till the end of the transient cooling process ( i-e reaching the steady state condition and τ_o=1 )in addition to its variation with distance from the centre.

The values are compared to that at the steady state condition An artificial neural network (ANN) is a computational model or a mathematical model based on biological neural networks.

It consists of an interconnected groups of artificial neurons and processes information using a connection approach to computation..

Artificial Neural Network using back propagation with transient and steady state has been used.

Five different ANN-models, using two algorithms: one for transient state type and steady state type.

Correlation was developed for the prediction of Nusselt number in stagnation region, along the distance from plate center region and average Nusselt number using the package MATLAB version (R2014a).

Developed correlations of Artificial Neural Network models are based on experimental data for the prediction of Nusselt number.

Results show that the Nusselt values are maximum at state of impingement process and decays with time on proceeding of the cooling process and also with space reaching the plate edge.