Failure Prediction (One of AI that can mostly expected to be adopted in Manufacturring)

Non-IT vendors (device vendors, plant constructors) IT vendors, and domestic manufacturers (with annual sales 10 billion yen or more)

I   What is Maintenance?

1. The way of thinking maintenance: Historical flow
2. from TBM (Time base Maintenance) to CBM (Condition base Maintenance)
3. To which facilities should CBM be adopted?

1. Priority of facilities
2. How to think about risks in this report

II 　 Failure Prediction Technology

1. Approaches from Viewpoint of Engineering and Data
2. Approach from Engineering Point of View

1. Target
2. Method
3. Judgement types

3. Approach from Viewpoint of Data

1. Statistical Method and AI
2. What is Machine Learning?  AI that automatically detects patterns and structures from data
3. Statistics vs Machine Learning?

4. PHM (Prognostics. & Health Management) and Failure Prediction

III   Limitation of AI

1. While Marketing encourages AI, Manufacturing being cautious about AI
2. Different backgrounds between Commerce and Manufacturing

1. At the time of anomaly detection, it is not clear whether the detection is the warning sign of failure or not (the issue of accuracy)
2. Though AI is able to predict failure, no next actions can be informed (the issue of cause and effect relationship)
3. The models are different individually, so that it is difficult to apply to similar things (the issue of individuality)

3. Barriers that prevent from making it into business
4. Is the limitation really the limitation?

IV   Forecast of Market Size

1. Market Environment of Failure Prediction + AI
2. Market Size and Future Transition

1. Current Status
2. Future Development
3. Future Scenario

V   Current Status of Maintenance

1. Compressors Being Primacy among Important Facilities
2. Considerable Number of CBM Utilization Expected for Important Facilities
3. Frequency of Maintenance
4. Measurement upon Failure
5. Necessary Information for Troubleshooting
6. Collecting and Utilization of Data
7. Unscheduled Maintenance Measurement Costs
8. Annoyance regarding Maintenance and Others
9. Accuracy level that may allow adopting failure prediction systems
10. IT Adopted at Factory Sites

VI   Case Studies (Trends of Attempts by Companies)

7 enterprises

• Data Collected
• Company Profiles
• Industry (details)
• Sales Size
• Factory Types

VII   Maintenance Status of Important Facilities

1. Facilities/Equipment Meeting Requirement
2. Maintenance Policy
3. Maintenance Frequency
4. Main Forms of Adoption
5. Failure Details, Measures at the time of Failure, Turning point of Maintenance
6. Necessary Information for Troubleshooting
7. Method to Collect Data
8. Data Utilization

VIII   Items regarding Factories as a Whole

1. Unscheduled Maintenance Measurement Costs
2. Annoyance regarding Maintenance and Others
3. Measurement against such Annoyance
4. Accuracy level that may allow adopting failure prediction systems
5. IT Adopted at Factory Sites