Why Should Manufacturers Adopt AI and Big Data?
Source – https://manufacturingglobal.com/
Manufacturing Global speaks to executive leaders at EY, Infor and GE Digital to get to the bottom of this question
Whilst the drive to digitally transform the manufacturing industry has been a topic of conversation for the last decade, recent events have only increased the need for the agility, scalability and resilience that Industry 4.0, smart manufacturing capabilities can provide. Speaking with Cobus Van Heerden, Senior Digital Product Manager at GE Digital, Mark Powell, Partner, EY (UKI Consulting), and Phil Lewis, Vice President, Solution Consulting EMEA at Infor Manufacturing Global looks at how technologies that harness AI and Big Data can help manufacturers unlock real-time operational visibility to achieve improved process reliability and performance.
What are the current applications of artificial intelligence (AI) and Big Data in the manufacturing industry?
CVH: Industrial AI uses a combination of targeted AI technologies, data, physics, and deep domain knowledge to solve key industrial business challenges. Traditional AI mimics human intelligence, whereas industrial AI builds upon it to unlock insights and determine causal knowledge in high-stakes, dynamic, and variable industrial environments. In Manufacturing, Industrial AI can be used to detect and predict key process and asset problems to help companies optimize their operations including capacity, quality, and cost structures.
PL: Textbook definitions of AI or Big Data miss the point that industries differ and will have drastically different demands for the technology. It is about the application of a given technology to a specific issue that a business may be experiencing. This issue may be an ‘industry-standard’ one or something that arises in the configuration of the technology. But there is the most value in the application of tools such as Big Data and AI to the critical 10% of a business that is truly idiosyncratic. We classify this as a 60/30/10 split and it is how we look to apply these technologies to drive maximum value.
For manufacturers looking to adopt Industry 4.0, smart manufacturing capabilities, why should manufacturers use AI and Big Data to do so?
CVH: Smart manufacturing deploys industrial advanced analytics to predict future asset and process performance using real-time and historical data and optimizing in a closed loop. This involves the use of AI and machine learning to enable process engineers to combine data across industrial data sources and rapidly identify problems, discover root causes of issues in the plant, predict the future performance of assets, and automate actions employees can take to improve quality, productivity, and operations.
MP: Digitisation is forcing manufacturers to reimagine their supply chains. As an example, most companies use internal data to track demand-supply balances and it is challenging for them to foresee external events impacting their supply chains. Using AI techniques that understand unstructured external data sets, such as social media and other data on events, manufacturers can plan for supply chain disruptions much sooner.
In addition, manufacturers can use AI and Big Data to build digital replicas of their manufacturing operations and tap into transformative possibilities of reducing cycle time in production, adding manufacturing capacity and predicting unplanned maintenance activities etc.
PL: Some of the poster child statistics for AI and Big Data simply demand attention. Recently, Siemens automated one of its factories in Germany, with 75% of the processes digitised or having increased automation. Productivity improved by 1,400%. That is game-changing for any business. This means many manufacturers are now looking at how they plug AI and Big Data into their plans for the future.
What is the best strategy for manufacturers striving to realise the value of AI and Big Data in their operations?
CVH: Process engineers have exceptional domain expertise to put together process models – or Process Digital Twins – and be able to interpret the models. This is the foundation for improving competitive advantage and success with analytics. To drive analytics and improve processes, manufacturers should put together a strategy that can align domain expertise to five capabilities: Analysis – automatic root cause identification accelerates continuous improvement; Monitoring – early warnings reduce downtime and waste; Prediction – proactive actions improve quality, stability, and reliability; Simulation – what-if simulations accelerate accurate decisions at a lower cost; and Optimization – optimal process setpoints improve throughput at acceptable quality by up to 10 per cent.
All process engineers can and need to develop capabilities in analytics and machine learning to remain competitive. Over time, engineers can go from small projects to pilots to multi-plant optimization with deep application of analytics. Their deep domain expertise provides a foundation for modelling processes and developing the analytics that are game changers in very specific applications.
Most importantly, get started with analytics. “Trystorm” some projects; put your intuitive ideas to the test and put data and analytics behind them. Don’t wait to become a data science expert. That isn’t necessary. Leverage proven easy-to-use industrial analytics tools fueled with your domain expertise. That’s going to drive big improvements quickly.
PL: Businesses – including manufacturers – tend to assess digital projects with a focus on either customer, supply chain, internal efficiency or people – those are the four main drivers for any foray into digital. These are often organic and arise from an ongoing ‘how can we do better’ attitude. This has been accelerated by concerns of competition as companies are now fearful of being left behind competition and disruptive entrants. There is palpable fear around being digitally relevant and this is promoting a lot of investment.
However, it is worth noting that many manufacturers have already invested heavily in technology (even before COVID forced a move to digitalisation) so the first point of definition is to align AI and Big Data to existing technology. When businesses assess their technology in use today, they need to bear in mind not only a short-term perspective of will the technology handle current processes but also does it provide a platform for the future? This latter perspective is built on data. Both elements are equally important but the second ‘platform perspective’ demands big data. It is no longer enough to choose a platform that just supports/tweaks the ongoing processes – there has to be future capabilities built-in.
There is then the need to ensure that this technology is deployed in the best way possible. This necessitates an open, cloud-based application landscape so a business can seize new opportunities such as Big Data or AI without having to go through a cumbersome integration and bolt-on process. This makes an organisation more agile, focusing on the creative application of the technology to the needs of the business, such as identifying new opportunities for revenue.
What are the challenges when it comes to adopting AI and Big Data analytics into manufacturing operations?
CVH: Manufacturers are challenged with reducing waste, costs, and risk while meeting customer demand. The combination of AI and data provides acceleration of digitisation through analytics-based solutions that empower workers with data in context so that people, assets, and processes work together efficiently.
Another challenge for companies is just getting started. They want to learn more about how to use analytics in their operations but don’t see it as a job for their current workforce. Fortunately, Industrial AI solutions can help and not require process engineers to be data scientists.
MP: The key challenge in adopting AI will come down to manufacturers’ ability to establish alignment across the organisation on some of the high-value areas where AI will make an impact. For example, using machine learning and computer vision to predict and identify faults in equipment before they occur, thus reducing production downtime and decreasing maintenance costs. Another challenge is establishing a culture of infusing AI into their processes through a test-and-learn culture.
For too long, organisations have talked about becoming ‘data driven’ and this has generally not worked as well as it had been hoped. Manufacturers need to take a different approach that starts with understanding where value can be driven from new insights and then focus on the data needed to drive the insights that can then drive business value. Organisations need to become ‘insight-driven and data enabled’ and not simply ‘data driven’ – only then will they really leverage the power of AI and big data.
PL: It is all about how attitudes towards data have changed. It was previously seen as a necessary evil but is now the number one asset in a business. Typically this drives an obsession with big data labels but it is what you do with the data that matters – using the likes of AI / BI / IoT etc to turn that data into a truly valuable asset. The automotive industry is the prime example – using and selling the data produced by a car. Interestingly, we now almost take ‘cloud’ for granted – had we answered this question 24 months ago, cloud would have been the first consideration, but it is now table stakes. It is no longer if a business will go cloud but more a question of what type of cloud/cloud use? – We have moved far beyond the infrastructure conversation –the how and into the what – and into the why a business looks to embrace digital.
Is artificial intelligence (AI) and Big Data driving the fourth industrial revolution (Industry 4.0)?
CVH: The combination of Industrial AI and data produces what we call a Process Digital Twin which helps manufacturers to rapidly troubleshoot continuous, discrete, or batch manufacturing process performance by mining insight from available sensor and production data. This technology, which utilises predictive analytics, enables users to analyse operating scenarios, qualifying the impact that operational changes will have on key performance metrics and identifying causes for performance variation. Digital Twins inspire continuous improvement, a key goal of the future of the industry by looking back to historical data as well as real-time to move forward rapidly.
PL: We see daily increases in AI/ML uses – inventory optimisation, maintenance, faster finance processes are all key areas that we see arise many times. For this to continue, and return on investment to continue, AI needs to be plumbed in and ready to go with other systems, rather than a bolt-on, or businesses face a hefty, and costly integration project. In terms of the next specific technology, it really depends on the maturity of the individual company or project – businesses are only just reaching the point of a digital fabric rather than a bunch of digital projects. Prescriptive working, driven by AI and fed by masses of sensor data, holds a huge amount of promise for the B2B / industrial markets and we see some very encouraging early shoots in asset maintenance and field service.