Tin tức

Tối ứu hóa đa mục tiêu gia công EDM bằng mô hình Kriging và thuật toán bầy đàn

Ngày đăng: 27-03-2017 / Đăng bởi: Khoa Cơ khí
Kết quả nghiên cứu này được đăng trên tạp chí: Materials and Manufacturing Processes, 10/2/2017. với tựa đề "Constrained multi-objective optimization of EDM process parameters using kriging model and particle swarm algorithm" http://www.tandfonline.com/doi/abs/10.1080/10426914.2017.1292037

 Tóm tắt:

This paper investigates the highly nonlinear relationship between process parameters and machining responses, including material removal rate, surface roughness, and electrode wear rate of electric discharge machining using Kriging model. Subsequently, an emerging multi-objective optimization algorithm called particle swarm was used to determine the best machining conditions that not only maximize the machining speed but also minimize electrode wear rate with a constraint of surface roughness. The experiment was carried out on CNC EDM machine with P20 steel (workpiece material) and copper (electrode). The research result shows that the material removal rate increases sharply when increasing the discharge current just like other researches pointed out. However, the relationship between electrode wear rate and current is complicated. Electrode wear rate appears the minimum value when the current is around 30 Amperes.The speed of change of material removal rate MRR per unit of electrode wear rate EWR is the highest when the surface roughness is around 14.5 um. The combination of Kriging regression model and particle swarm optimization is considered as an intelligent process modeling and optimization of EDM. Proper selection of process parameters helps the EDM operator to reduce the machining time and cost.

Kết luận:

This paper proposes the Kriging model to express the relationships between EDM process parameters and machining responses for P20 steel and then optimizes the process parameters using particle optimization method. The main conclusions from the investigation of this work can be drawn as follows:  

1. The relationship between process parameters and electrode wear rate is highly nonlinear. The second-order response surface methodology cannot be used in this case. As the result, Kriging is the versatile alternative model. There is extremum point of electrode wear rate in the process window. The lowest electrode wear rate appears when the current rate is around 7.5 (equivalent to 30 Amperes).

2. The speed of change of material removal rate MRR per unit of electrode wear rate EWR is the highest when the surface roughness is around 14.5 m.

3. The optimization problem in which material removal rate and electrode wear rate are objective functions, and the surface roughness is a constraint, is a general optimization problem that is more practical and realistic than single objective or three or more objectives in EDM machining. 

4. It is difficult to state that how many percent of improvement of objective function obtained by optimization compared to un-optimal case. The answer depends on a certain machining case. However, the best process parameter can only be obtained by optimization. Proper selection of process parameters obviously reduces the machining time so as to lower the EDM machining cost.