원문정보
초록
영어
Wood plastic composites (WPCs) were prepared and used as pencil board. The primary raw materials were corn straw fiber powder, high-density polyethylene, and polystyrene. The tensile strength, surface roughness, hardness, and roll cutting performance of the prepared composites were tested and compared with those of linden wood, which is typically used for pencil board. A back propagation (BP) neural network was adopted to build the prediction model of process technical parameters of the composite. An improved particle swarm (modified particle swarm optimizer, MPSO) was adopted to optimize the BP neural network, and the advantages of the PSO algorithm’s global optimization ability and the BP neural network algorithm’s high processing speed were realized. The results show that the composite had the best performance, with a mass ratio of 3.8 (corn straw fiber powder): 47.6 (high-density polyethylene): 28.6 (polystyrene). The composites can be used as a substitute for linden wood in the production of pencil board.
목차
1. Introduction
2. Pencil Composite Plate BP Neural Network Modeling and Algorithms
2.1. BP Neural Network Model
2.2. BP Neural Network Algorithm Analysis
3. Improved Particle Swarm Optimization and MPSO-BP Neural Network Hybrid Algorithm
3.1. Decreasing Nonlinear Inertia Weight Strategy
3.2. Asynchronous Changes in the Value of the Acceleration Factor
4. Experiment Research
4.1. Design Variables
4.2. Process Temperature
5. Result and Discussion
References