Determinants of Modulus of Rupture and Modulus of Elasticity of Particleboards on the basis of Data base

The MOE and MOR are controlled by production variables of particleboard process. Now، the basic question is which of the particleboard variables is effective on bending strength property? 13 variables of internal scientific resources were measured with 100 repeats. The study steps include the following; liner regression or stepwise، Genetic algorithm، and Artificial Neural Network. The number of effective variables was selected from the output of the stepwise procedure and then modeling of these variables using WinGamma and Mathlab. The number of effective variables in the modulus of elasticity and modulus of rupture are equal to 10 and 9، respectively. As the results of Gamma test shows، effective variables of bending strength are as following: 1. Type of wooden raw material، 2. specific gravity of raw material by simple averaging، 3. specific gravity of raw material by weighted averaging، 4. UF percent، 5. particleboard density، 6،7، and 8. Time، tempreture، and pressure of press. In the other hand، particleboard thickness of MOR and also، percent mixture of wood and non-wood lignocellulosic materials and NH4CL percent for MOE are important. Results of Genetic algorithm and Neural Network were showed that some variables can be kept constant and particleboard properties are controlled by these effective variables، but specific gravity of raw material by weighted averaging for MOR and UF percent of MOE have the strongest effect. Result of BFGS Neural Network has shown that mean absolute percent error of MOR and MOE are equal 5. 644% and 4. 91%، respectively.