基于遗传算法的激光切割镍基合金质量优化 下载: 917次
Quality Optimization of Laser-Cutted Ni-Based Alloys Based on Genetic Algorithm
1 大连海事大学航运经济与管理学院, 辽宁 大连 116026
2 辽宁科技大学应用技术学院, 辽宁 鞍山 114051
图 & 表
图 1. 激光切割示意图
Fig. 1. Schematic of laser-cutting process
下载图片 查看原文
图 2. 实验宏观图
Fig. 2. Exp erimental macrograph
下载图片 查看原文
图 3. 样本金相数据的标记图。(a) 2号;(b) 3号;(c) 4号;(d) 5号;(e) 6号;(f) 7号;(g) 8号;(h) 9号;(i) 10号;(j) 11号;(k) 12号;(l) 14号;(m) 15号;(n) 16号;(o) 17号;(p) 18号;(q) 19号;(r) 20号;(s) 21号;(t) 22号;(u) 23号;(v) 24号;(w) 25号
Fig. 3. Mark graph of metallographic data of samples. (a) No. 2; (b) No. 3; (c) No. 4; (d) No. 5; (e) No. 6; (f) No. 7; (g) No. 8; (h) No. 9; (i) No. 10; (j) No. 11; (k) No. 12; (l) No. 14; (m) No. 15; (n) No. 16; (o) No. 17; (p) No. 18; (q) No. 19; (r) No. 20; (s) No. 21; (t) No. 22; (u) No. 23; (v) No. 24; (w) No. 25
下载图片 查看原文
图 4. 17~20号样本的4次取样图。(a) 17号1次样本;(b) 17号2次样本;(c) 17号3次样本;(d) 17号4次样本;(e) 18号1次样本;(f) 18号2次样本;(g) 18号3次样本;(h) 18号4次样本;(i) 19号1次样本;(j) 19号2次样本;(k) 19号3次样本;(l) 19号4次样本;(m) 20号1次样本;(n) 20号2次样本;(o) 20号3次样本;(p) 20号4次样本
Fig. 4. Four-time sampling figures of samples from No. 17 to No. 20. (a) No. 17, first sampling; (b) No. 17, second sampling; (c) No. 17, third sampling; (d) No. 17, fourth sampling; (e) No. 18, first sampling; (f) No. 18, second sampling; (g) No. 18, third sampling; (h) No. 18, fourth sampling; (i) No. 19, first sampling; (j) No. 19, second c sampling; (k) No. 19, third sampling; (l) No. 19, fourth sampling; (m) No. 20, first sampling; (n) No. 20, second sampling; (o) No. 20, third sampling; (p) No. 20
下载图片 查看原文
图 5. 三层BP网络
Fig. 5. Three-layer BP network
下载图片 查看原文
图 6. BP神经网络的预测结果。(a)预测值与期望值对比;(b)误差;(c)误差百分比
Fig. 6. Resutls predicted by BP neural network. (a) Comparison between predicted value and expected value; (b) error; (c) percentage of error
下载图片 查看原文
图 7. 算法流程
Fig. 7. Flow chart of algorithm
下载图片 查看原文
图 8. 适应度曲线
Fig. 8. Fitness value curve
下载图片 查看原文
图 9. 验证实验图。(a)正面宏观图;(b)背面宏观图;(c)金相微观图
Fig. 9. Experimental diagram for test. (a) Positive macrograph; (b) back macrograph; (c) metallographic micrograph
下载图片 查看原文
表 1GH3128的化学成分(质量分数,%)
Table1. Chemical compositions of GH3128 (mass fraction,%)
Composition | Ni | Cr | W | Mo | Al | Ti | Fe | B | Zr | Ce |
---|
Value | Bal. | 19.0-22.0 | 7.5-9.0 | 7.5-9.0 | 0.4-0.8 | 0.4-0.8 | 1.0 | 0.005 | 0.04 | 0.05 |
|
查看原文
表 2因素水平
Table2. Factor levels
Symbol | Factor | Level 1 | Level 2 | Level 3 | Level 4 | Level 0 |
---|
A | Electric current /A | 200 | 210 | 220 | 230 | 215 | B | Pulse width /ms | 0.8 | 1 | 1.2 | 1.4 | 1.1 | C | Cutting speed /(mm·min-1) | 150 | 200 | 250 | 300 | 225 | D | Defocusing amount /mm | -1 | -0.5 | 0.5 | 1 | 0 |
|
查看原文
表 3样本综合评分
Table3. Comprehensive scores of samples
No. | Level ofABCD | S /μm | K /μm | l/L /% | Sc | No. | Level ofABCD | S /μm | K/μm | l/L /% | Sc |
---|
1 | 1111 | 0 | 0 | 0 | 0 | 14 | 4231 | 170 | 171.5 | 84 | 70.85 | 2 | 1222 | 155 | 177.5 | 100 | 84.5 | 15 | 4324 | 272.5 | 193 | 100 | 77.08 | 3 | 1333 | 260 | 195 | 100 | 77.5 | 16 | 4413 | 417.5 | 209.5 | 100 | 68.18 | 4 | 1444 | 262.5 | 200 | 100 | 76.88 | 17 | 0000 | 242.4 | 148.875 | 100 | 82.99 | 5 | 2123 | 157.5 | 159.5 | 100 | 86.18 | 18 | 2222 | 169.9 | 178.625 | 100 | 83.64 | 6 | 2214 | 250 | 224 | 100 | 75.1 | 19 | 3333 | 213 | 183.375 | 100 | 81.01 | 7 | 2341 | 197.5 | 171.5 | 100 | 82.98 | 20 | 4444 | 310.5 | 204.75 | 100 | 74 | 8 | 2432 | 102.5 | 262 | 100 | 78.68 | 21 | 1000 | 207.5 | 162 | 100 | 83.43 | 9 | 3134 | 132.5 | 109.5 | 80 | 73.94 | 22 | 2111 | 190 | 161.5 | 81 | 68.32 | 10 | 3243 | 172.5 | 155 | 100 | 85.88 | 23 | 3222 | 220 | 128.5 | 100 | 86.15 | 11 | 3312 | 175 | 219 | 100 | 79.35 | 24 | 4333 | 165 | 183.5 | 100 | 83.4 | 12 | 3421 | 235 | 273.5 | 100 | 70.9 | 25 | 0444 | 247.5 | 195 | 100 | 78.13 | 13 | 4142 | 0 | 0 | 0 | 0 | | | | | | |
|
查看原文
表 4样本误差
Table4. Sample errors
No. | Sample | S /μm | K /μm | Sc | Error /% | Average |
---|
17 | 1 | 328 | 155 | 78.1 | -6.26 | | 82.99 | 2 | 211.5 | 162 | 83.225 | 0.28 | | | 3 | 192.5 | 133 | 87.075 | 4.69 | | | 4 | 237.5 | 145.5 | 83.575 | 0.69 | | | 18 | 1 | 179 | 181 | 82.95 | -0.83 | | 83.64 | 2 | 188 | 186 | 82 | -2 | | | 3 | 157.5 | 166.5 | 85.475 | 2.15 | | | 4 | 155 | 181 | 84.15 | 0.61 | | | 19 | 1 | 221.5 | 176 | 81.325 | 0.39 | | 81.01 | 2 | 221.5 | 176.5 | 81.275 | 0.33 | | | 3 | 221.5 | 181 | 80.825 | -0.23 | | | 4 | 187.5 | 200 | 80.625 | -0.48 | | | 20 | 1 | 326.5 | 190.5 | 74.625 | 0.84 | | 74 | 2 | 283.5 | 219 | 73.925 | -0.1 | | | 3 | 322.5 | 209.5 | 72.925 | -1.47 | | | 4 | 309.5 | 200 | 74.525 | 0.7 | | |
|
查看原文
表 5隐层节点误差
Table5. Hidden layer node errors
Number of nodes | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 |
---|
Error | 31.89 | 15.62 | 19.95 | 16.53 | 18.86 | 17.66 | 25.12 | 19.47 | 11.26 | 37.33 | 57.09 |
|
查看原文
张艺赢, 曹妍, 陈宇翔, 牟向伟. 基于遗传算法的激光切割镍基合金质量优化[J]. 激光与光电子学进展, 2018, 55(11): 111404. Yiying Zhang, Yan Cao, Yuxiang Chen, Xiangwei Mu. Quality Optimization of Laser-Cutted Ni-Based Alloys Based on Genetic Algorithm[J]. Laser & Optoelectronics Progress, 2018, 55(11): 111404.