본문 바로가기 주메뉴 바로가기
22대 대한민국 국회 국회정보길라잡이 국회의원검색
로그인 English 국회부산도서관

전체메뉴

HOME

정보검색

소장정보 검색

국회도서관 홈으로 정보검색 소장정보 검색
결과 내 검색 동의어 포함
결과 내 검색 동의어 포함

-

목차보기

Title Page

ABSTRACT

Contents

NOMENCLATURE 12

1. Introduction 13

1.1. Research background and motivation 13

1.2. Research objective 13

2. Literature reviews 15

2.1. The history of overall chatter detection method 15

2.2. The limitation of existing method 15

2.3. Outline of this work 16

3. Experimental investigation on chatter detection and real-time monitoring system using a mobile machine tool 17

3.1. Stability lobe diagram and vibration simulation for customized mobile machine tool 17

3.2. Training data 18

3.2.1. Machining conditions 19

3.2.2. Signal to noise ratio compensation 19

3.3. Feature Extraction 22

3.3.1. Analyzing periodic characteristics 22

3.3.2. Statistical feature 23

3.3.3. T-test 25

3.3.4. Validity of feature extraction 27

3.4. Deep neural network 31

3.4.1. The definition and characteristics of deep neural network 31

3.4.2. Structure of deep neural network 32

3.4.3. Hyperparameters Configuration 32

3.5. Deep neural network training 34

3.6. Result 35

3.6.1. Validation of simulations 35

3.6.2. Validation of experimental vibration 37

3.7. Conclusion 40

4. Developments of real-time monitoring system 41

4.1. Establishment of wireless communication 41

4.1.1. Experimental setup 41

4.1.2. Construction of TCP/IP communication 41

4.2. Development of application using Android studio 42

4.2.1. Algorithm development for transmitted data preprocessing 42

4.2.2. Chart implementation using MPAndroidChart 43

4.2.3. Model embedding 45

5. Concluding remarks 48

5.1. Overall Conclusions 48

5.1. Path forward 48

REFERENCES 49

List of Tables

Table 3-1. The machining parameters of experimental training data 20

Table 3-2. The statistical features for deep neural network training 24

Table 3-3. The top 50 feature extracted through the T-test 26

Table 3-4. The machining conditions for experimental validation 38

List of Figures

Figure 1-1. The schematic diagram of research objective 14

Figure 3-1. The overall mobile machine tool setup 17

Figure 3-2. The stability lobe diagram of the mobile machine tool 18

Figure 3-3. Experimental setup for slot milling process using mobile machine tool 20

Figure 3-4. The illustrated representation of the sliding window 21

Figure 3-5. The illustrated depiction of the process involving wavelet decomposition 22

Figure 3-6. The coverage of frequency ranges related to decomposition levels 23

Figure 3-7. Entire feature extraction process 25

Figure 3-8. The configuration of a two-sided test 26

Figure 3-9. Number of features based on decomposition levels 27

Figure 3-10. The waveform of vibration simulation: (a) axial depth: 0.18 mm, (b) 0.20 mm, (c) 0.24 mm, (d) 0.26 mm 29

Figure 3-11. The distribution of vibration simulation: (a) axial depth: 0.18 mm, (b) 0.20 mm, (c) 0.24 mm, (d) 0.26 mm 30

Figure 3-12. The FFT amplitude of vibration simulation: (a) axial depth: 0.18 mm, (b) 0.20 mm, (c) 0.24 mm, (d) 0.26 mm 31

Figure 3-13. The illustrated representation of the deep neural network 32

Figure 3-14. The illustration of (a) ReLU, (b) Leaky ReLU (c) Sigmoid 33

Figure 3-15. Training and validation data (a) loss, (b) accuracy 35

Figure 3-16. Validation accuracy of vibration simulation 36

Figure 3-17. 3D scatter plot of prediction values for (a) 1100 rpm, (b) 1975 rpm, (c) 3850 rpm 37

Figure 3-18. The validation on the experimental data for (a) N＝3, (b) N＝2 (c) N＝1 40

Figure 4-1. The schematic diagram of TCP/IP communication 42

Figure 4-2. Problems of the TCP/IP communication 43

Figure 4-3. Preprocessed transmitted data 43

Figure 4-4. Initial version of application user interface 44

Figure 4-5. User interface design using adobe XD 44

Figure 4-6. User interface design including chart implementation 45

Figure 4-7. (a) Model embedding method and (b) stability evaluation user interface 46

Figure 4-8. Real-time monitoring application using tablet PC 46

초록보기

 The Mobile Machine Tool (MMT) represents a transformative technology in machining, embodying characteristics such as mobility, flexibility that set it apart from traditional machine tools. Despite its advantages, the mechanical movements of MMT can introduce variations in Material Removing Rate (MRR), leading to undesirable vibrations known as chatter during machining operations. The timely detection of chatter is crucial for mitigating several consequential issues. Chatter can adversely affect the quality of machined products by causing surface irregularities, compromising finish, durability, corrosion resistance, and aesthetics. Additionally, chatter accelerates cutting tool wear, necessitating frequent tool changes, resulting in elevated downtime, increased costs, and reduced productivity. The deviation from intended tool paths can introduce inaccuracies in finished products, further contributing to rejections and added expenses. Furthermore, chatter-induced reductions in cutting speed and feed rate diminish machining efficiency, elongating cycle times, and decreasing throughput. Ultimately, detecting chatter is crucial as it increases the overall cost in the manufacturing process.

The primary objective of this research is to address the intricate nonlinear nature of chatter through the application of deep learning methods. Leveraging the capability of deep learning to capture nuanced nonlinear relationships inherent in machining data, our study aims to develop a robust chatter detection model. Additionally, we explore the integration of smart devices for real-time monitoring by developing an application embedded with the trained deep learning model. By achieving effective chatter detection through deep learning and real-time monitoring with smart devices, this research seeks to enhance machining quality, improve productivity reduce overall costs. The proposed approach holds promise for preventing the adverse effects of chatter, thereby contributing to the realization of high-quality products and efficient manufacturing processes using MMT.

추천서가 (다양한 추천 자료를 만나보세요)

권호기사보기

권호기사 목록 테이블로 기사명, 저자명, 페이지, 원문, 기사목차 순으로 되어있습니다.
기사명 저자명 페이지 원문 기사목차