Driving predictability and minimizing variation are the holy grails of manufacturing. Today, technology plays a significant role in that pursuit with manufacturing becoming increasingly sophisticated. Five trends have emerged:
1. A proliferation of data collection devices
2. A new set of data storage decisions
3. The use of more robust data and a keen need for analytics
4. More automation and computer integrated manufacturing
5. Increased complexity and demand on networking
Data Collection Devices
There has been a proliferation of data collection devices in the manufacturing industry in recent years. Sensors throughout the production line and on a wide variety of transport equipment bring great value– from minimizing machine downtime to preventing product defects and misdirection.
Sensors can provide information to be analyzed in real-time and accessed by plant operators or others as needed to make better decisions. By establishing parameters to detect abnormalities in the manufacturing process, alarms, such as text messages, can be generated automatically to notify plant personnel of an anomaly that needs to be addressed in real time.
The data also can be valuable in retrospect: if a customer experiences a problem with our flooring, even long after it’s manufactured, the data we collect allows us to conduct cause-and-effect analysis to make any necessary changes to avoid similar problems in the future.
This vast data collection prompts key questions for all manufacturers: What data do we keep? How much do we store and for how long? Where do you store it? If a device captures a data point every microsecond, is that necessary? Or would a data point every second be just as effective?
The answers, of course, depend on your operations. Making those decisions is a collaborative process between IT and manufacturing teams. Over simplified, the most critical first step is to agree upon what the company wants to achieve with its data, what’s valuable real-time, and what will be valuable to analyze over a longer period and in conjunction with other data—acknowledging those parameters will inevitably change over time.
While data storage costs have improved with the use of new data storage technologies such as Hadoop and cloud-based archives, storing unnecessary data can become cumbersome and expensive. Finite resources (i.e. budget) are often a helpful prioritization tool. And no discussion of cloud storage would be complete without careful consideration of information security.
As noted, determining how information will be analyzed is critically important. Data is more complex than ever. Information is no longer just words and numbers. Alphanumeric information is now gathered alongside visual and audio data. Combining data from a variety of sources to get an actionable picture requires new thinking, new methods, and new tools.
How do you manipulate what you store? How do we collapse it into consumable form? What will be most valuable to different parts of the organization? It’s easy to fall into the trap of “we might need that someday,” but experience shows that the propensity to keep and/or report too much can result in analysis paralysis as well as greater-than-necessary costs.
Much larger volumes and varieties of data drive the need to use new tools to conduct the analysis and to provide interpretations of the result. The market is rich with options and zeroing in on finite choices is now proving to be difficult.
Automation allows manufacturers to control processes in real time. Factors such as pressure or humidity can be controlled to prevent variation that leads to defects. At the same time, advancements in robotics are allowing machines to complete tasks that consistently produce high quality goods. The combination of process control and robotics helps us make the most of our resources and reduce waste, but also improves efficiency, cost and safety.
With more than one billion dollars invested in new equipment, technology and processes over the past few years, Shaw’s operations are more complex than ever. As a result, almost every job at Shaw—from designers and data analysts to machinists and managers—requires a higher skill level than in the past.High-tech operations bring sustainability improvements, require more highly skilled workforce and create better paying jobs.
With new sources of data at higher volume comes the question of how to effectively route traffic. Numerous control endpoints need to talk directly with each other, in real time, within a network often saturated with traffic. Data also must be routed to multiple targets to enable tiers of visibility into operations. High levels of detail are necessary for control systems, granular summary data is needed for machine-level quality control decisions, and larger-grained summaries are used for factory optimization needs.
With all this moving data traditional routing approaches using hardware quickly become expensive in scale. Software-Defined Networking (SDN) makes the dynamic nature of today's applications more manageable without requiring physically-enabled routes. Software-based routing drives cost and complexity out of the system.
As new technology emerges, these five trends will continue to evolve for the foreseeable future.
Consider that more than half of all IP traffic will originate with non-PC devices by 2018, according to the Cisco Visual Networking Index, with Machine-to-Machine (M2M) modules seeing an 84 percent increase from 2013 to 2018, representing more than 7 billion M2M connections, across an array of industries including manufacturing. Further, GE has estimated that the new wave of innovation brought about by what it calls the “Industrial Internet” could boost global GDP by as much as $10-15 trillion over the next 20 years, through accelerated productivity growth.
Whether you call it a smart, intelligent or brilliant factory, advanced manufacturing is here, and IT professionals in this sector should be prepared for continued data deluge resulting from the exponential growth in the four V's of Big Data: volume, velocity, variety and veracity.