Unlocking the power of AI in manufacturing

We explore what AI-powered technology can bring to the manufacturing industry and how data is key to unlocking its potential

Artificial Intelligence (AI) is starting to be implemented in almost every aspect of everyday life and is becoming increasingly important in industry operations. As the world continues to digitally transform, AI is evolving and becoming a key driver in digitalisation. With the ability to automate what would otherwise be a costly process, AI can be a game-changer at all levels of the value chain - this is particularly significant in the manufacturing industry.

AI-powered technologies, for manufacturers, can allow for direct automation, predictive maintenance, reduced downtime, 24/7 production, improved safety, reduced operational costs and greater efficiency.

John Spooner, Head of AI, EMEA at H2O.ai shared his insights into the benefits of AI in manufacturing, he said: “A successful project must have a positive Return on Investment (ROI). The creation of artificial intelligence to drive automation allows projects to remove slow and tedious processes by moving from experimentation to production. As a result, the high cost of resources such as people and infrastructure can be minimised by making them more productive and efficient.”

Ultimately, AI gives businesses more control over their manufacturing process. As it delivers insights on capabilities and capacities, manufacturers are able to more easily identify the best way to obtain the maximum performance from their assets. 

As with all technologies, there is a caveat to this. Many organisations and manufacturers, particularly those who have smaller operations, cannot afford the high implementation and maintenance costs that come with these complex technologies. Adding to this, the introduction of new, AI-powered technologies removes the human element to processes that have typically been carried out manually. Humans can offer reasoning to support their conclusions, AI and machine learning is a ‘black box’ that typically cannot explain or defend its answers.

However many leaders believe that the benefits outweigh these costs, Ramon Antelo, VP, Digital Manufacturing & Operations Business Development, Capgemini Engineering highlighting this, he said: “Many organisations are at the cusp of digital transformation and the pandemic has fast-tracked the need to digitise their current infrastructure and this is high on their priority list.” 

“Once the digital transformation is complete, AI will play a critical role in creating a robust manufacturing environment that can be remotely enabled and monitored,” he continued. 

Understanding AI to successfully improve manufacturing operations

Undoubtedly, the uptake and implementation of AI have accelerated in recent years. According to Antelo, this acceleration has come from the original equipment manufacturers (OEMs) as there is a push to edge computing in the industry. He adds: “This [edge computing] allows AI algorithms to run very close to the machines without a significant investment in Industrial Internet of Things (IIoT) or cloud infrastructure.”

The abundance of data manufacturers can obtain has also allowed for increased adoption of this technology, which relies on data to create efficient algorithms. Spooners explains: “Due to the advancements in machine learning, all of this data can now be utilised to create intelligence about what is and isn’t working in the end to end manufacturing process.  Ease of accessibility and the rapid decrease of cost of the “computational power” needed to run the machine learning algorithms on this amount of data has helped increase the adoption of this technology.”

“Data is the main ingredient that helps manufacturers utilise AI’s power – whether that be sensor data, vibration data, audio data or image data,” added Spooner.

It is key, however, that when looking to implement AI technologies manufacturers understand how to effectively scale this technology, to do this there needs to be a level of expertise and knowledge around the best practices when implementing AI. Expanding on this, Antelo said: “Many companies have tried unsuccessfully to incorporate AI in their manufacturing process and get stuck in “pilot purgatory” where they are unable to scale their projects. Some early AI adopters have managed to deploy machine learning algorithms, but many are stymied by scaling them.”

Data as a key enabler for AI-powered technologies in manufacturing

With data being such a key enabler of AI in manufacturing, it comes as no surprise that digital twins can support manufacturers as they look to adopt AI technology. 

Digital twins, a virtual representation that serves as the real-time digital counterpart of a physical object or process, originated from NASA who used it as a practical definition to improve physical model simulation of spacecraft in 2010. The US government agency used basic twinning ideas for space programming by creating physically duplicated systems at ground level to match the systems in space. 

Now, digital twins can support manufacturers as they can provide more information to the data scientists and IT teams in manufacturing organisations.

These experts can use the digital twins for trial and error. By applying AI to the simulation data that is generated from these virtual replicas manufacturers can identify problems before they occur. Additionally, if manufacturers integrate this data with their physical counterparts they can gain further insight into their AI systems.

As already mentioned, to implement AI successfully, there needs to be an understanding of AI technologies as well as the ability to understand how to leverage the vast amounts of data an AI-enabled manufacturing process generates. This challenge remains when manufacturers look to create digital twins to support their operations. Highlighting this issue, Antelo said: “Building a digital twin ecosystem is complex and requires the expertise to manage and interpret vast amounts of data from an AI-enabled manufacturing process.” 

“Implementing digital twins without an understanding can significantly slow down your manufacturing process. It is critical for the organisation to understand its requirements and what it hopes to achieve with the implementation of digital twins before it embarks on its strategy. AI in manufacturing can deliver a wealth of benefits and digital twins is one of many tools organisations can reap the benefits from,” he added.

If these challenges are tackled and the manufacturers have obtained a sufficient understanding of AI processes and digital twin management, Spooner explained how the benefits are significant: “The advantage that digital twins bring is that we can optimally create these processes, in the first instance, based on the combination of AI and the simulated data. With the additional wealth of data generated by the digital twin, systems can be created that self-diagnose and self-correct with very little intervention from a human.  AI, in conjunction with digital twins, will allow manufacturing to take the next step in automation."


Featured Articles

Andrew Ng Joins Amazon Board to Support Enterprise AI

In the wake of Andrew Ng being appointed Amazon's Board of Directors, we consider his career from education towards artificial general intelligence (AGI)

GPT-4 Turbo: OpenAI Enhances ChatGPT AI Model for Developers

OpenAI announces updates for its GPT-4 Turbo model to improve efficiencies for AI developers and to remain competitive in a changing business landscape

Meta Launches AI Tools to Protect Against Online Image Abuse

Tech giant Meta has unveiled a range of new AI tools to filter out unwanted images via its Instagram platform and is working to thwart threat actors

Microsoft in Japan: Investing in AI Skills to Boost Future

Cloud & Infrastructure

Microsoft to Open New Hub to Advance State-of-the-Art AI

AI Strategy

SAP Continues to Develop its Enterprise AI Cloud Strategy

AI Applications