Chris Barlow explores the shift from traditional perceptions of AI to its real-world benefits in manufacturing, focusing on creating a data-driven strategy to allow AI tools to support operational goals.
Every manufacturing solution, big and small, is about something more than machines or even the data captured from those machines; instead, every solution should serve the people working in your operation, placing their needs and capabilities at the forefront.
The road to value from AI for your workforce and operation is paved by developing a data-driven business. Being “data-driven” is an idea that’s often oversimplified into building dashboards and KPIs. But those kinds of solution only “look backwards” – and in limited ways. Instead, with a clear strategy and structured data, it’s possible to re-use and repackage data from assets and processes to inform your operational activities. Such as:
- Machine/process data helps evolve planning and process standards, supplier qualification processes and more
- Delivering good SOPs in operational areas where machines can’t provide all the answers
- Supplementing review/analysis with decision support tools
- Getting the most out of your overall systems footprint (ERP, MES, Inventory, etc) using freely flowing plant data
Creating a data strategy in manufacturing involves several critical steps designed to align data management with business objectives. Starting at the end point, the ideal ways of working for your operation, allows you to develop your data strategy with a clear goal in mind. By establishing a vision, you can do the groundwork for data to support a myriad of business needs from day one.
But we can’t simply buy digitalisation from a vendor, as lovely as that would be! So here are our 6 steps to take when developing a manufacturing data strategy:
- Define Objectives: How do you want data to support your business and drive decision-making? Business goals such as improving operational efficiency, enhancing product quality, or driving innovation are great places to start.
- Outline Your “Future” Data Framework: Industrial data is often “messy” – machines and sensors that do the same things may present their data with different structures and details. Fortunately, technology vendors offer several ways to allow you to create consistent outputs and add other useful information so a range of uses can be supported. This can be done purely through “edge processing,” mapping automation data to simple asset/enterprise models within SCADA and Historian frameworks, or to fuller event models (typical in a manufacturing execution system/MES). Decisions made here will also affect how you develop data governance policies to ensure quality, security, and compliance.
- Prioritise Initiatives: Identify and prioritise data initiatives based on their potential impact and feasibility. This could include implementing advanced analytics, predictive maintenance or IoT integration.
- Assess Current State: How mature is your data collection? Do you have any connectivity gaps? Evaluating your existing data infrastructure is the first step in developing your approach to data. It involves understanding what data is collected, its structure, how it is stored, and how accessible it is.
- Technology Selection: Choose appropriate technologies and tools that align with your data strategy. This may involve investing in new data platforms, analytics tools, or upgrading existing systems. Different tools will have their own focus and strengths, as such it’s likely you’ll consider one kind of AI for schedule optimisation, a different one for speeding process optimisation, a different one altogether for understanding reliability issues.
- Implementation Plan: Develop a phased implementation plan with clear milestones and deliverables. Ensure alignment with business operations to minimise disruption.
At the core of any data strategy are people. From the top floor to the shop floor, data empowers operators with decision support, continuous improvement, personal development and business growth.
At last, we come to the hot topic – AI.
Good data empowers AI in the same way as everyone in your operation. However, AI doesn’t need to be treated as special.
When a business is data-driven, AI tools, like Proficy CSense, can interpret information across your operation to provide insights using advanced analytics and machine learning. These tools enable users to detect anomalies, predict failures, and optimise performance by tapping into the complex streams of data that would be challenging for humans to harness in real-time.
So, how do you go from strategy to reality? What other benefits, efficiencies and capabilities can you unlock with a data-driven business?
Find out exactly how we can turn a strategy into a shop-floor reality with our manufacturing digitalisation guidebook.
Chris Barlow is the Technical Director at Novotek Solutions UK & Ireland.
With nearly 30 years’ experience, Chris shapes manufacturing solutions that support key business strategies such as manufacturing agility, right-first-time programs, and sustainable manufacturing.
For more articles like this, visit our Industrial Data & AI channel.
