20 If your company is ready to drive forth, here are some foundation questions and answers to set your course: Q&A FOR UNDERSTANDING AI AND HOW TO GET STARTED Q: In addition to the obvious machine wear/tear alerts, what other value can AI add? A: Eliminate day-to-day operational setbacks and be able to highlight actionable insights for C-levels using engaging charts and graphs. As an AI deployment creates a holistic data-based view of the entire foundry process, it often reveals — and helps remedy — problems with current methods. One example is a customer’s quality system that reports completely different batch volumes compared to the rest of the line. That makes it impossible to build a working process model but is usually easily resolved. AI presents data in real-time with KPI dashboards to visually show and alert the operator to important indications that can prevent breakdowns, reduce energy consumption or other targets that are set by the operator. Understanding fully how to implement AI-suggested parameters is another vital enabler. Which machine settings do operators change and by how much to reach the ‘AI control zone’? Make sure the operators know how to achieve this. To make it happen, an AI vendor who completely understands foundry machinery and processes is essential. To understand exactly what is happening with its process and sensecheck what the AI is recommending, the foundry must give its expert users full access to process data — and the tools to report on and analyze it. But an IIoT cloud database can make process and all other data linked to it visible (ERP, quality, maintenance) to any permitted user, so think about where else that data could deliver added value. Monitoring and alerting on the line are obvious applications but you could also think about business intelligence reporting for your CFO. Q: To get started, who should be on our project team? A: A dedicated team is essential to plan your AI project and drive it forward. To do this successfully, it must have the right members. This includes a committed, executive-level sponsor with the authority to overcome internal obstacles to change. Factory floor supervisors and representatives of the machine operators who will actually implement the AI’s recommendations must also be closely involved. Foundry-wide user compliance with AI recommendations for machine and process settings is absolutely central to success. As a general guideline, compliance rates below 80% make it almost impossible to accurately link the effect of AI recommendations to process outcome — and that breaks the feedback loop that optimization depends on. It is essential that the project team promote AI user adoption. Q: How do I engage our lean team with additional duties? A: Promote success for each operator with giving them independent dashboard control. Run regular project team meetings and update all the operators involved on current results that prove the benefit of compliance with AI prescriptions and what the next set of goals are. Announcing better scrap results, publicizing upcoming training, repeatedly emphasizing the changes required–all these reinforce good behaviors to make them personal and permanent. Give each operator a screen by their machine with their own dashboard that shows the target range of, for example, melt temperature or sand moisture level suggested by the AI. Then operators can pick their own settings required to reach that operating zone. Also, regularly show operators how better control of their own sub-process contributes to overall improvements in scrap. Show that the vision is being realized, thanks to their efforts and publicize the team’s overall contributions. Managers also need reports (again, delivered through simple dashboards) showing compliance rates for each
RkJQdWJsaXNoZXIy NDI4Njg=