Manufacturing and the fourth revolution
Increasing issues such as volatile resource prices mean that manufacturers need a more agile approach allowing them to adapt to changing conditions, pursue profit opportunities and minimize risk. On top of this, an upsurge in the Internet of Things (IoT) throughout the business has begun to trigger a transformation. IoT is a network of physical objects—devices, equipment, engines and robotics—which use sensors to collect and exchange data.
What’s next for manufacturing?
Industry leaders within manufacturing have already set in motion the idea of a fourth industrial revolution, or Industrie 4.0. Building on the introduction of robotics and information technology (IT) in the early 2000s, this is the next step in "smart" manufacturing, which is focused on utilizing data and connectivity for business value.
Through the IoT, manufacturers will eventually be able to interlink every part of their business—equipment, systems, services, and even human activities—so that one automatically communicates with another to inform decision making within the value chain, both internally and externally. Yet as this comes into full fruition, manufacturers will be faced with granular information from multiple locations and sources, causing a data explosion to occur.
It’s the power of this connectivity that the fourth industrial revolution puts at its heart. Yet, just adding new sensors, telemetry and smart devices will not deliver full business value. The challenge is in making use of the large volumes of data generated, and for that manufacturers must embrace machine learning—a new phase in advanced data analytics.
Better business outcomes
With today’s digitalization manufacturers can now supplement typical data collection, from production lines and technical log processes, with granular information from telemetry, sensors, and all types of machine-generated data—providing manufacturers with enormous prospects for data-driven optimization.
Yet, too many manufacturers continue to stop at the water’s edge and do not use their data to its full potential. Some, for example, will use the data to track a product’s journey on the manufacturing line, but not to improve the operations on this journey.
Machine learning is a complementary technology used to analyze large datasets, identify underlying patterns, action prescribed outcomes and actively learn from the outcomes to feed future automated decision-making. All to a volume and speed that humans simply cannot envisage, much less perform. Machine learning’s greatest advantage over previous forms of data analytics is that it does not require deep understanding of the technology itself, nor to have data interpretation capabilities.
For example, when applied to equipment maintenance, an essential, often unplanned activity that can easily disrupt the workflow, machine learning can have a great effect on costs and productivity. Imagine being able to accurately predict equipment failure before it occurs, and action "just-in-time" maintenance. By applying machine learning to both historical and real-time equipment and output data, the manufacturer is able to identity signs of maintenance needs before the faults manifest themselves and automatically make the appropriate service recommendation. This allows manufacturers to improve operations, drive greater productivity and improve competitiveness.
While the algorithms used will change to suit the data and company’s purpose, this machine learning can also be put to other uses, like demand and load predictions, optimizing the supply chain and monitoring production processes with the help of computer vision. This data-driven automation and intelligence brings manufacturers one bold step closer to the agile business model required for Industrie 4.0.
Leading the revolution
Big data analytics and machine learning technologies, first applied vastly inside internet and online businesses, are now being applied to almost all industries. The truth is, manufacturing is the only industry best placed to lead the revolution, with faster and more efficient adoption. Why? The industry’s existing culture.
Where a marketer from a bank or an offline retail company would struggle accepting the need to work with black box predictions, run correct A/B tests and measure every action, manufacturing is already one step ahead. Manufacturing has a long established practice of working by numbers, integrating new scientific developments into old processes, running experiments and comparisons, and valuing the importance of optimization.
While many people are still skeptical about calling this a revolution, data-orientated technologies are doubtless driving greater efficiency, higher production and faster fulfillment, which equal new opportunities. What manufacturer can afford to ignore that?
Jane Zavalishina is CEO of Yandex Data Factory. This article originally appeared at Internet of Business. Internet of Business is a CFE Media content partner. Edited by Chris Vavra, production editor, CFE Media, firstname.lastname@example.org.
See related stories linked below.