How Democratized Artificial Intelligence Can Move Manufacturing to a New Evolution Pace
correlation complexity and few device / sensors data consolidation due to small data storage size and computing power. Even if those technologies are innovating fast, the other issue that presents itself is the innovation upgrade( mostly hardware) which is limited by the complexity of deployment and maintenance at a large scale. New devices like smart watches, tablets, AR glasses, collaborative robots and 3D printers are now also able to extend their capabilities with a direct access to the cloud and hybrid cloud, using external computing power and services.
2. In case of brownfield or low computing power capability in the devices, this intelligence can be hosted in the edge computing devices, which are playing the role of automation controller( PLC, Programming Logic Controller 24 / Motion Logic Controller). They are now becoming intelligent controllers with a mix of industrial real-time functions,“ human real-time” smart computing and cloud gateway capability to access external cloud computing power and services. This technology can also help to control and make smarter a group of machines and devices to automate actions and monitor a part of a manufacturing process in a specific location area. The limitations here are the silo scope and the limited local storage capability and analytics computing power.
3. The third level is the plant level, consolidating data from factory machines, humans, orders, warehouses, logistics, maintenance and quality to better organize the activities using cloud or hybrid cloud AI( hybrid cloud means hosting a part of the cloud on premise). The aim of this level is to leverage complex analytics to provide predictive alerts, machine learning and adaptive intelligence to help technicians, workers, managers and operational officers make the best decision in advance or in realtime to improve overall efficiency and excellence. As a plant manager, I want to predict in real-time my future production achievement, my supply, logistic, worker and maintenance plans based on the current situation and predicted situation, and I want to be agile if the situation changes. That is what Cloud / Hybrid Cloud AI is providing.
4. Finally, the last layer is the ecosystem, regional or worldwide level, consolidating data from your company, getting access to data from suppliers, benchmark data from partners, consulting firm and data suppliers to refine your horizontal and vertical AI analytics to make the best decision. Can I benchmark predictive failure of all my motors across my factories and OEM suppliers to identify specific( humidity, amperage fluctuation, vibration, etc.) patterns of failure in my conveyors, robot and elevators and avoid hard failure? How can I predict the impact of
24 https:// en. wikipedia. org / wiki / Programmable _ logic _ controller
- 16- November 2017