I often drawn parallels between our industry and the TV show "How It's Made," explaining that we, control systems engineers, function behind those fascinating scenes, choreographing the intricate ballet of industrial technology, processes and people. As part of this, we often speak about how data is a valuable byproduct of the control systems we deploy.
Historically, the data consumers we worked with were at the plant level. But we have recently experienced a deeper level of interest among our customers in enterprise-wide data accessibility.
The group of data people involved in “How It’s Made” has grown to include IT systems architects, data scientists, network/security engineers, operations technology (OT) subject matter experts, quality analysts and production supervisors. Among these groups, a commonality we are seeing is the goal of having one central, contextualized source of truth for production-related data for their entire organization.
This is an industrial data ops approach, and it is becoming abundantly clear that our traditional choreography needs some revisions.
What are industrial data ops?
Industrial data operations are about bringing together system architecture, data access methods and cultural approaches in the way industries manage and utilize their data. It's about creating synergies, adding context and governance to data, and ensuring that various data sources across an enterprise merge fluidly, driving optimized results with shortened data analytics cycles. Industrial data ops are enabled by deploying a common data platform, which allows for data modeling, contextualization and data consumption across multiple applications, and doing all of this at scale.
Historically, our realm of focus consisted of integrating the PLC and HMI to deliver immediate access to process data for analysis by site maintenance, engineering, production, quality and continuous improvement personnel. Just providing a means to store historical data with the ability to build trends to their hearts’ content felt like a job well done. But as industries are becoming more digitally interconnected, simply storing and trending historical data isn't enough.
Now, we face more intricate questions as other personas across the enterprise want access to production data. This reality raises questions such as:
- Where and how is the data directed post-historization?
- Who is leveraging this data and for what purpose?
- What other data sources can add valuable context to the existing data?
- How do various applications use the data?
- How is the data secured and governed?
The data management landscape
On the OT front, there is still a considerable reliance on purpose-built historians and analytical toolsets. These systems have been tailored over the years to meet the unique needs of OT environments, providing real-time insights and optimizing on-site operations. However, when we shift our gaze to the corporate spectrum, the narrative changes.
Analysts at this level are increasingly advocating for a cloud-first approach. Their reasons are clear—cloud infrastructures offer scalability, flexibility and compatibility with an array of big data toolsets. For them, it's not just about understanding what happened in the past, but also about predicting future trends, optimizing supply chains and deriving strategic business insights.
In a recent project, we found ourselves navigating a complex landscape of diverse data sources and passionate consumers of data within a manufacturing organization. The goal was to foster an industrial data ops approach that would seamlessly bring together data from production, quality analysis, research and development, and engineering. On paper, it seemed straightforward, but it involved straddling multiple data worlds, each with its own intricacies.
While production leveraged a cloud-based MES system, QA and R&D combined spreadsheets, a homegrown LIMS (laboratory information management system) and historian data. Engineering employed traditional historian data with HMI trending tools, but they also wanted to overlay lab data, specifically manual data entries from the LIMS system. To harmonize these disparate data sources, we crafted a common data model independent of its origin. One of the most valuable aspects to this deployment was introducing event data (production stages, steps and alarms) adding crucial context to time-series data. While each department still maintains their distinct tools, the enterprise now benefits from a unified, enriched data landscape driving more efficient analysis and decision making.
Redefining manufacturing for the digital age
So, while I'll continue to draw analogies with "How It's Made," the current narrative has added layers. It's no longer just about controls engineering choreography but also about orchestrating a grand symphony of data. As control systems engineers, our roles are expanding, and as we embrace these changes, we're poised to be an integral part of our customers’ data strategies in a way we never imagined before. Where once our expertise was confined to the plant floor, we're now diving into the abstract realms of data analytics, cloud integration and cross-functional collaboration.
This evolution not only enhances our technical proficiencies but also enriches the value we bring to the table. We're not just focused on efficient operation of production equipment; we're making sure that data works for the business. As we step into this expansive role, we are shaping the future of the industrial landscape, one contextualized data tag at a time.
Dan Malyszko is vice president at Malisko Engineering, a certified member of the Control System Integrators Association (CSIA). See Malisko Engineering’s profile on the CSIA Industrial Automation Exchange.
