Select Page

By Tim Burke and Jennifer Davis

It is time for manufacturers to demand more from their data, and it is time for them to take control of that data.

This is a golden age of efficient, responsive manufacturing. With tough competition and little room for error, today’s manufacturers continue to grow and remain the backbone of the world’s economies, accounting for an estimated one-sixth of global GDP with the top 5 countries alone contributing a combined $8.6T in manufacturing output.

But they have a problem. 

While manufacturing data can be incredibly valuable, it’s still highly siloed and not fully utilized. Even with advancements in manufacturing execution systems (MES) and machine vendor software systems, the data within manufacturing environments remains separate and, without a solution tying them all together, will remain this way for the foreseeable future.

It’s certainly not for lack of effort. Manufacturers have gone to great lengths to gather and contextualize their data. It’s simply that accessing and using the siloed data, even with the assistance of vendor software and programming interfaces, is still always a project– a project that takes time and resources and is therefore focused on a single question or problem, limiting both its scope and its value.

Just as factories have standardized processes for a long time, we see that it is time for the global enterprise to better standardize how it collects and utilizes machine data. The factory can no longer be a collection of expensive data siloes, each with a low utilization warehouse. Put simply, that practice threatens the livelihood of the entire business.

Industry 4.0 is about competitiveness

Manufacturing is a starkly competitive environment. Practically overnight a factory down the block (or in another hemisphere in our global economy) can innovate and out-compete on capability, performance, or price. Each manufacturer uses the best technology available to keep ahead. The next revolution, that of the much-expected “Industry 4.0,” is about integrating that data regardless of the source. The company’s data collection and analysis systems must partner equally with all vendors and tools within the organization so the company can have visibility that enables proactive improvements across both the factory floor and the global enterprise.

All signs and trends indicate that manufacturing is on the fast path to adopting new technologies for automation, visibility, and strategy.

To be successful, these technologies need to tie the data from across the organization into a meaningful picture to wrest the most value from these new capabilities. There are significant challenges to doing this, but the tools are here. The solutions to finally bring together contextualized data at scale exist, and they work phenomenally well.

Proprietary data, locally stored inside the factory, such as output of a machine vendor’s software, is a data silo. Arch’s vision of Industry 4.0 is to bridge all these proprietary data sources and do it continuously; operational questions should not necessitate months-long projects. They need instantly actionable insights.

While individual machine vendors are making strides in the collection and application of data, they are not providing this holistic answer. 

Most provide software that adds value to their machines and makes them run better. Often it is designed standalone, with minimal external linkages for a line or factory. This design simplifies IT needs for day-to-day factory operations but hinders the ability to look holistically at data across the organization. Realistically, all manufacturers are using machines from multiple manufacturers. They need to look forward- using data in context. While locally deployed vendor software solutions are valuable to the factory, they still contribute to, rather than solve the data silo problem.

As compared to the constant, varied “data projects” manufacturers have become accustomed to, the process of collecting and contextualizing machine data can, in fact, be automated. Once connected, the data is simply readily available to quickly answer any question. Arch works with manufacturers worldwide, connecting and analyzing complex machine data sets. ArchFX is inherently agnostic to the various machine types. It extracts all the rich data (not just summaries or rolled-up statistics), structures it in a useful manner, and contextualizes it for real-world use.

Apples to apples: Context is everything

A large contract manufacturer believed it was operating at peak efficiency. However, a deeper analysis of global machine data revealed that they were losing $30 million a year of unplanned downtime from ten top errors that happened across all their factories.

Now that powerful tools like ISA95 guide enterprise manufacturing to integrate control and operations, there are continual revelations unfolding. Organizations are comparing data across factories and timelines to create new, enhanced best practices.

To do this, data needs to be more complete. The information needs to be in context so that the comparisons are alike: comparing apples to apples. If one batch of data is from a line down for maintenance, another is from a new product introduction (NPI), and another is from high volume production, comparing them creates confusion, not clarity.

Contextualization means three things.

Putting data from each machine into its place in the ISA95 reference architecture so you have a complete picture of a line, area, and site with data from all machines.
Joining data between machines to find new insights. For example, can you combine PNP recipe data with the related AOI defect data? If so, defects are not just known by their reference designators but by the exact feeder, nozzle, and head that placed it. It could be (and often is) the case, that problems with a single PNP nozzle are causing all of the line’s defects.  Until you join the data together though, you can’t see the root cause with clarity.
Joining in the factory intent. Advanced data analysis techniques can spot the patterns of mass production vs. NPI vs. non-production maintenance in the rich machine data so when you aggregate information across the enterprise, you are able to compare apples to apples. This can leverage operator data entry, although machine learning algorithms can also be used without any manual entry.

Let a solution anywhere become a solution everywhere

Contextualized data is actionable data. As we mentioned before, this can be (and sometimes is) deployed today using months-long historic data projects and investigation. However, the new competitive landscape won’t wait. Data needs to be contextualized instantly and be readily available, as well as preserved in its full richness for historical analysis at any point in the future. 

Common problems that are occurring in multiple factories are opportunities for massive ROI gains. With the data at hand, the full attention of the most valuable experts can be focused on these high-value questions.

Similarly, in the apples-to-apples theme, if one factory is experiencing a problem that others are not, the high functioning lines can inform the problematic one of the proper fix. When the team in charge of the problem line has visibility into the other operations, the fix can be employed quickly.

Informed alerts are the cherry on top

Global analysis of errors from one PNP vendor showed that a few specific error codes would lead to long line downtime if unaddressed. Quick action would prevent the outage. From the global analysis, an alerting system now notifies the responsible engineer when the error occurs. The engineer receives a link to the solution playbook to resolve the error

While getting the data is no small task, even as valuable as it is, it’s not the best part of what can be done with data once it’s un-siloed. Inherent to the platform is an alert system. Those alerts fuel the ArchFX action management system. A no-code solution allows alerts and the dashboard to target specific behaviors.

When contextualized data crumbles data silos, problems are acted upon quicker and better. Some of these are long-standing problems that were invisible when data was still siloed. No project needs to be budgeted for, no “I’ll get back to you with that.” Questions are asked and answered. 

Visibility across the organization elevates the idea of “best practice” to not just a local practice, but a global one.

Are you ready to see the value that’s trapped on your factory floor? Visit https://archsys.ionow to schedule a demo of the ArchFX platform.