At the ongoing Virtual AI Summit in Hong Kong, Abhinav Singhal, Chief Strategy & Innovation Officer, Thyssenkrupp, spoke about data-driven insights amplifying human ingenuity and transforming shopfloor.
AI has huge potential in manufacturing. There are things like inbound logistics, R&D, production, distribution and sales, and customer service. You are talking about over 500TB of data getting generated in a plant. There is route and network optimization, automated warehouse and picking, spend analytics, advanced analytics by product design, design engineering and visualization, advanced process control and yield management, predictive maintenance and asset management, digital twins for remote monitoring and QC, etc. There is need for user interfaces and software, analytics, storage and cloud, etc.
AI has not been adopted that strongly in manufacturing. This is very typical of manufacturing. For them, reliability, productivity and quality matter more. There is legacy IT, as well. There may be machines from different suppliers. The frontline buy-in is another factor. AI can enrich their work, and make them more productive.
There is need to adopt a modular approach to implement AI in the plant. There are data selection layer, connectivity layer, and storage, analytics and AI layer, respectively. There should be a portfolio of options. There should be the tools to choose from. You can accelerate the scale up of the AI initiatives.
Next, there should be focus on predictive and prescriptive use cases. There should be aggregation of unstructured and structured data from multiple sources. There is a need to identify anomalies. There should be predictive models for forecasting and prescriptive failure diagnosis learning algorithms. There should be data-driven recommendations on optimized actions and strategies.
As you move to predictive and prescriptive use cases, you start getting the maximum out of AI. The neural network predicts multiple parameters of the resulting product properties. You are able to run a system at the optimal product quality. There are many examples of using predictive and prescriptive tools that can boost AI and provide optimal production.
AI will not replace humans. Humans who can work with AI will replace those who cannot work. Human intelligence, along with digital expertise, will create AI for your operations. There is need to reskill the frontline. Focus on predictive and prescriptive use cases for the highest impact.