AI and IIoT: How to make products better and faster – at less cost (Reader Forum)
Artificial intelligence (AI) is a key asset in the continuing effort to derive greater value from business processes and business models in general.
While much of the attention AI has garnered has centered on digital services and applications, the technology is also driving significant enhancements to industrial and manufacturing environments.
Like their counterparts in the data center, industrial executives and operators face a daily challenge to do more with less and devise more efficient and profitable ways to convert raw resources into finished products.
Naturally, the more data that can be drawn from these processes, the more efficient and effective they become. But this is only true if this data can be compiled and analyzed in meaningful ways.
AI fulfills this need, primarily due to its singular ability to parse large volumes of unstructured data without burdening data scientists and analysts with the complex and time-consuming process of preparing data for analysis.
Much of the data AI leverages in a factory setting is sourced from the myriad connected devices that contribute to the industrial internet of things (IIoT).
Similar to its consumer-facing cousin, IIoT is designed to allow intelligent management systems to peer into industrial infrastructure at a granular level, delivering crucial data on system health, manufacturing processes, resource consumption and a host of other factors.
Ultimately, this data is compiled and parsed to provide insight into how products can be made better, faster, cheaper and more relevant to the needs of consumers.
One of the ways AI contributes to a more effective IIoT environment is its ability to coalesce data from multiple stores and multiple formats. In any given setting, industrial processes are overseen by multiple distinct software platforms, such as SCADA, ERP and CMMS, which are generally hosted on a mix of ageing, on-premises infrastructure.
Multiply this by numerous facilities across regional, national or even international footprints, and you get an idea of the challenge facing today’s manufacturing entities – even those that have already embraced a high level of automation.
These disparate data systems often leave “blind spots” in industrial processes that are major contributors to waste and inefficiency. When analysts lack a cohesive view of the end-to-end industrial process, they not only fail to recognize the emergence of known deficiencies that disrupt that process but are unaware of the unknown issues that arise as well.
For example, a monitoring system overseeing an initial step may not be delivering optimal results for a secondary step, which in turn leads to a mismatch in final assembly – not enough to make the product unusable but perhaps to cause an early failure and dissatisfied customers.
Lacking the proper data into their industrial processes, many executives resort to the time-honored tradition of on-the-job experience (aka, a hunch) or employ rigid assessment procedures to inform their decision-making.
These approaches often lead to misjudgments that result in wasteful spending, if not outright disruption of the manufacturing process. An example is the practice of scheduled maintenance. Many organizations pull equipment off-line for replacement or refurbishment based on nothing more than vague estimates of wear and tear over a specified time period.
No one really knows whether the system or a given component is actually near failure, but it’s better to replace it now than to take the risk, even if it leads to inefficiency in the manufacturing process or cost overruns.
AI offers the ability to identify and correct these issues, forming a crucial bridge between data and business metrics. With an AI-driven IIoT platform in place, management systems gain real-time data into the full manufacturing process, right down to individual components on the assembly line.
For manufacturers today, this level of data analysis not only prevents failure but continually optimizes the process to improve quality and lower costs. In today’s economy, even minute gains in efficiency and waste reduction can spell the difference between success and failure, particularly as economies of scale kick in.
All of this helps industrial plants become more proactive instead of reactive. Managers can head off real problems as they develop rather than throw money at problems they may not even have.
And as the machinery itself becomes more flexible, we can expect intelligent, autonomous systems to fine-tune themselves to achieve a continuous state of optimization and efficiency. At the same time, plants will be better able to craft new products tailored to specific customer demands, all of which will drive increased brand loyalty and market differentiation.
It has been said that every business these days is a digital business. Regardless of what product or service is being delivered, data drives the entire business model – from initial design and manufacture, to marketing, sales, delivery and support.
Businesses that better understand their data will be the ones to thrive in the new digital economy.
But perhaps the most profound effect of industrial AI is not on the machines or the processes AI guides, but on the human workforce, which will finally be able to shed the drudgery of data management and routine maintenance to do what it does best – imagine, create and innovate.
Prateek Joshi is an artificial intelligence researcher, an author of eight published books, and a TEDx speaker. He is chief executive of Plutoshift, which provides asset performance management for industrial process facilities to increase energy efficiency while maximizing throughput.