IIC Journal of Innovation 6th Edition - Page 10

How Democratized Artificial Intelligence Can Move Manufacturing to a New Evolution Pace Figure 1: Move your data to the cloud for large scale multi sources of data and intensive AI and ML computing then use the outcomes locally to improve your business behaviors in your specific context specific issues and potentially reach the target of autonomous system maturity level. anticipate the alert of a potential failure of a system within a timeframe – such as within the next 15 minutes or 10 hours. In fact, the next data capture disruption wave can potentially be the blockchain to get a massive capture of transactions, contract and ownership behaviors data. That gives information to the asset manager or a machine itself to take a counter-action decision in advance. What changed from previous reporting and forecasting method? Now, we are showing potential futures in real-time. Leveraging the new computing power capability, we are now able to process a large source of data streamed together, correlate them in real-time, compare with historical data, then forecast and show the future, all of that in real-time. Of course, we can apply this on predictive maintenance, but also in production trends, sales trends, financial trends and many other individual or consolidated indicators. Another case could be, you are a plant manager, and you want to know in real- time your actuals and your potential production trends for the next hours. F OUR M AJOR AI T OPICS ARE K EY FOR M ANUFACTURING C OMPANIES 1. The predictive analytics (or time machine) is now starting to be embedded in business applications (predictive maintenance, production monitoring, demand & production & supply planning and financial planning). This approach is using a small or medium set of historical data mixed with current data, correlated together in a business context to define a potential future situation against a timeline. Let’s take the case of predictive maintenance: The wish is not to get an alert when the system fails, but IIC Journal of Innovation - 9 -