Martin Lehmköster

Transcription

1 Place for, company logo from speaker Reduction of Downtime, Quality Improvement and Customer Satisfaction with High Speed Web Inspection Systems Martin Lehmköster

2 Agenda 1. Introduction to ISRA 2. Reduce costs, increase productivity 3. Basics of web inspection cameras light GUI 4. Classification higher ROI 5. DataMining make money out of your data 6. Rewind Manager to reduce costs 7. Validation Martin Lehmköster 2

3 Agenda 1. Introduction to ISRA 2. Reduce costs, increase productivity 3. Basics of web inspection cameras light GUI 4. Classification higher ROI 5. DataMining make money out of your data 6. Rewind Manager to reduce costs 7. Validation Martin Lehmköster 3

4 ISRA worldwide LANSING HERTEN (Germany) web inspection center SHANGHAI ATLANTA DARMSTADT ISRA AG TAOYUAN world leader in web inspection more than 3500 systems installed 280 employees Martin Lehmköster 4

5 Agenda 1. Introduction to ISRA 2. Reduce costs, increase productivity -examples 3. Basics of web inspection cameras light GUI 4. Classification higher ROI 5. DataMining make money out of your data 6. Rewind Manager to reduce costs 7. Validation Martin Lehmköster 5

6 Situation We are facing the fact of increasing production costs due to: Increasing Raw Material Costs Increasing Energy Costs More critical Customers At lower market prices for the products on global markets! Martin Lehmköster 6

7 Avoid costs of machine downtime detected defect in paper (hole and contamination) caused machine stop in the print machine Martin Lehmköster 7

8 Web Inspection Finds your defects as early as possible to document and classify the defects to take corrective action to avoid more defects to avoid processing scrap and low quality to avoid customers complaints Martin Lehmköster 8

9 Reduce costs of quality change in raw material Blown film, transparent, tubular film, transmission Width: 400mm, v = 80m/min Change of pellet at 9800m Martin Lehmköster 9

10 Reduce costs of quality change in raw material Blown film, transparent, tubular film, transmission Width: 400mm, v = 80m/min Change of pellet at 9800m Martin Lehmköster 10

11 Uncritical defects PVC, calender, transparent, transmission, v=50m/min Martin Lehmköster 11

12 Critical defects PVC, calender, transparent, transmission, v=50m/min Martin Lehmköster 12

13 Agenda 1. Introduction to ISRA 2. Reduce costs, increase productivity 3. Basics of web inspection cameras light GUI 4. Classification higher ROI 5. DataMining make money out of your data 6. Rewind Manager to reduce costs 7. Validation Martin Lehmköster 13

14 System-Architecture Cameras Position Encoder Illumination Cabin Data base Operating unit Martin Lehmköster 14

15 System overview Data base Display of inspection data Reports Alarm signals Interface to networks Martin Lehmköster Statistic 15

16 System for diaper film/ nonwovens Martin Lehmköster 16

17 System for paper Cameras: 8 (2.048 pixel) Machine: PM#25 Width: mm Speed: 1200 m/min Illumination: LED white Resolution: MD: 0.71 mm CD: 0.38 mm Rewind-Manager Marker: Ryeco 6 color Martin Lehmköster 17

18 Light sources fluorescent light - cheap - simple maintenance LED - free of maintenance Special development??? Reflection Transmission Martin Lehmköster 18

19 Ultrabright LED Light Line lifespan: 10 years! Martin Lehmköster 19

20 Many pixels for the right decision Web inspection can handle 320 Megapixel / sec per camera Martin Lehmköster 20

21 Resolution and Speed resolution Speed of the camera is the key for high quality images! Martin Lehmköster 21

22 Large Images on Demand Dielines in transparent film of 250µ Martin Lehmköster 22

23 Advantages More pixel better resolution More pixel better pictures More pixel better defect detection More pixel better defect classification Martin Lehmköster 23

24 User verification High Resolution Images = Accurate Classification Identify Defect Cause & Take Corrective Action to Eliminate Defect Martin Lehmköster 24

25 Agenda 1. Introduction to ISRA 2. Reduce costs, increase productivity 3. Basics of web inspection 4. Classification higher ROI 5. DataMining make money out of your data 6. Rewind Manager to reduce costs 7. Validation Martin Lehmköster 25

26 Criteria for a simple/ single classification? Criteria: Box - Classifier - bright / dark -area - lenght / width equal? Martin Lehmköster 26

27 A Learning System for Classification A A A A A A letter = A B B letter B = B Martin Lehmköster 27

28 B Classification with Quick Teach Reliable classification Z V S D S I = this is the letter A B W B () () E Ü P... even with noisy background! Martin Lehmköster 28

29 Requirements for classification High accuracy Simple operation with images Display of human classifier - high amount of criteria Easy modification/ high flexibility Real time ability/ immediately response Intuitive operation Martin Lehmköster 29

30 Box-classifier Only for simple criteria: - area, length, width Simple manual definition Unqualified for complex tasks Danger of forgotten! Division of defects according to simple criteria Pseudo defects! Very good and stable solution for simple task should be included in every system as a plain classifier Martin Lehmköster 30

31 Today: Next neighbour -classification Self-learning classifier Mathematics is hidden Easy modification Little basic information (Images) necessary No over-teaching possible Anytime extendible Avoiding pseudo-defects A fast learning classifier gives a higher ROI Martin Lehmköster 31

32 Fast learning classifier! results? Get your production run at once with a classifier time - do not waste time with teaching a classifier Martin Lehmköster 32

33 User-friendly learning The user selects a defect out of the continuous production Easy to understand, easy to operate Guarantee for quick implementation Description Saving of information back to data base of system Martin Lehmköster 33

34 Learning Requirements Classifier needs image gallery For every defect class images must be available Images will be collected during production automatically Martin Lehmköster 34

35 Agenda 1. Introduction to ISRA 2. Reduce costs, increase productivity 3. Basics of web inspection 4. Classification higher ROI 5. DataMining make money out of your data 6. Rewind Manager to reduce costs 7. Validation Martin Lehmköster 35

36 Data Preparation reducing data to a reasonable extent Defectdatabase + list Time table Individual report Martin Lehmköster 36

37 Example: Single lane report to rate quality Martin Lehmköster 37

38 STEP III global network central data mining Martin Lehmköster 38

39 Agenda 1. Introduction to ISRA 2. Reduce costs, increase productivity 3. Basics of web inspection 4. Classification higher ROI 5. DataMining make money out of your data 6. Rewind Manager to reduce costs 7. Validation Martin Lehmköster 39

40 Marking defects Martin Lehmköster 40

41 Winder Interface & Control Main Inspection System Configuration Data Defect Data Rewind Manager Rewind Controls PLC Martin Lehmköster 41

42 Defect Marking and Web Position Synchronization Start Defect Mark (Bottom Side) Web Sync Marks (Top Side) ½ Martin Lehmköster 42

43 Rewind Manager total system solution Rewind Manager to automatically stop re-reeler for defect removal Defect marked at defect location on top surface Codes marked on bottom surface Mark Reader controls reel map display to show image and location of next defect to stop on Operator override to skip defect Rewind Manager Sending slow down pulse 2000 ft before defect PLC Decode Barcode PLC Break Control Martin Lehmköster 43

44 Rewind Manager Screen - Map Defect filter Current Position Next defect on the web Martin Lehmköster 44

45 Rewind Manager - Defect Selection Stop on Defects in Red Click on a defect image to toggle stop status Martin Lehmköster 45

46 Agenda 1. Introduction to ISRA 2. Reduce costs, increase productivity 3. Basics of web inspection 4. Classification higher ROI 5. DataMining make money out of your data 6. Rewind Manager to reduce costs 7. Validation Martin Lehmköster 46

47 Validation by artificial defects Image of an artificial defect Martin Lehmköster 47

48 Validation by real defects Electronic masking of real defect images Signal Signal Martin Lehmköster 48

49 Advantages of high-end web inspection Reduction of - customer complaints -of waste -of downtime - of costs Deliver proven quality High return on investment Martin Lehmköster 49

50 Thank you for your attention! Thank you for your attention! Questions? Martin Lehmköster ISRA Surface Vision GmbH Albert-Einstein-Allee Herten Tel Martin Lehmköster 50