Skip to content

Leveraging Digital Analysis for the IIoT, Big Data

Pat Waurzyniak
By Patrick Waurzyniak Contributing Editor, SME Media
Jon Sobel, CEO and Co-Founder of Sight Machine Inc. in San Francisco

Jon Sobel, CEO and Co-Founder of Sight Machine Inc. in San Francisco, sat down with Senior Editor Patrick Waurzyniak to discuss numerous topics related to Controls and Software, including how Sight Macine Inc. uses analytics software to help shops leverage plant-floor data.

Manufacturing Engineering: Manufacturing systems are getting more connected than ever. Describe how your company’s analytics software helps shops leverage their plant-floor data.

Jon Sobel: Manufacturing generates enormous amounts of digital data, but until now manufacturers have struggled to turn that data into useful insights. At first glance, you might assume that data volume is the principal challenge. It’s not. The main issue is more subtle—it is the variety. Modern factories generate hundreds of data types: from sensors, PLCs, cameras, historians, serialization, databases, and MES, SCADA and ERP. The data is often collected and stored, but until factories can overcome the variety problem, it’s not useful.

How do you put this raw data into context to help improve operations and quality? That’s what Sight Machine does. People increasingly refer to IIoT [Industrial Internet of Things] and Big Data as being like the five senses for industry: in the same way that sight, smell, and touch help us as people make sense of a constantly changing, nuanced world, the right digital technologies can help industry understand their operations with a level of rigor and speed that’s never before been possible.

Sight Machine’s solution takes digital data of just about every type and continuously analyzes that data. There are four steps: we acquire the data from factory floor sources through connectors and adapters we’ve written; we then clean, condition and integrate the data into manufacturing models that enable meaningful analysis; we run the data through our analytics; and we visualize the results via Web pages available on any phone, tablet or laptop. This provides a real-time view into operations. Manufacturers are using Sight Machine’s platform for a variety of applications including real-time monitoring of production and quality; retrospective analysis of product variation and failures; predictive analytics to support process improvement; and benchmarking on multiple levels—contractor, factory, line and tool.

ME: What are some new features in your software?

Sobel: We’ve spent four years building out a systematic way to integrate and analyze factory data; now that the platform is built, we’re going deeper into analysis. We’ve recently added root cause analytics and the ability to continuously relate any production parameter to any outcomes of interest—basically the ability to continuously regress everything in the plant against everything else. This is the foundation for anomaly detection and advanced predictive modeling. We’ve also built real-time alerts for situations where process drifts, and we’re developing materials management tools that use production data to help reduce waste and improve sustainability. More is on the way.

ME: Big Data and the IIoT are getting a huge amount of hype. How will they help manufacturers over time?

Sobel: IIoT and Big Data are buzzwords for sure—but the underlying notion could not be more simple. Information is valuable. If you can know more about how to design, build and support your products, you are going to have an edge. Going digital helps companies run their businesses better, improve relationships with customers, and innovate faster. This is real.

ME: How does your software integrate/analyze data that you’ve gathered from the plant floor from PLCs, sensors and IoT-enabled devices?

Sobel: Since many factory data sources work off of common protocols [Modbus, OPC, Profinet, etc.], we are able to reuse and scale this library. Our solution uses Linux, Spark, Python, R, and leading front-end Web technologies to move data into a modern data store and then analyze and present that data through Web services. We also work with data residing on data historians.

Sight Machine uses both local and cloud computing resources and can work with private or public clouds. This is a different approach from traditional software; we’re a Web company. One symbolic way to think about this is what would it be like to have Google for your factory: it’s one thing to query a database, it’s another to have deep intelligence that’s relating and surfacing the data for you in useful, relevant ways.

Because this is fundamentally a data solution, not some sort of legacy software, it is highly compatible with existing systems and technologies like MES or ERP, along with newer tools such as Hadoop. We don’t replace anything, and our customers don’t have to remove or change anything to use us. Our platform is open and agnostic. As you would expect, we have APIs [application programming interfaces] on our platform, so we can make data available and enable application development on top of our service. This makes us compatible with all major systems.

ME: What are the most important shop-floor metrics, including OEE [Overall Equipment Effectiveness], for analyzing manufacturing productivity?

Sobel: In general, OEE is an excellent overall framework for analysis, if it is calculated rigorously and consistently across the enterprise—a much more difficult task than might be expected. It is also imperative that those using OEE go beyond dashboards and counting into analysis. Once underlying data is validated, manufacturers can move from KPIs [key performance indicators]—what did I produce today—to analytics. Why is this equipment performing poorly? When can I expect it to fail? What’s causing our quality issues?

ME: What industries stand to gain the most from using your technology?

Sobel: Digital transformation is most likely in those industries where supply chains are complex, quality is imperative, and regulation requires demonstrable control over extended operations. These include automotive, pharmaceutical, food and beverage, medical device, oil and gas manufacturing, electronics, and defense and aerospace.

ME: Who are some key customers using your software?

Sobel: An automotive OEM is using Sight Machine to understand and improve a process that spans several factories: powertrain parts are made in one plant, machined in a second, and finished in a third. By combining process data from one plant with quality and process data from the other two, we’re helping them understand how to prevent costly defects.

We are assisting a large consumer manufacturer with the deployment of new automation in 15 of its factories. Given the scale of this manufacturer’s equipment rollout, it is imperative that they develop standardized best practices. They need data and analytics to do that. In these cases, we’re taking existing data to help manufacturers solve critical problems. This is an exciting time for manufacturing.


Siemens AG (Munich) announced Nov. 25 that it will acquire Polarion Software GmbH (Stuttgart, Germany), developer of a browser-based application lifecycle management (ALM) enterprise solution. The acquisition will further enhance Siemens’ support for systems-driven product development, a holistic development approach that combines systems engineering with an integrated product definition, in an open environment. Siemens will add Polarion’s offerings to its Teamcenter PLM software, making ALM an integral part of its product development process. The transaction is expected to close during the first quarter of 2016. No financial details of the acquisition were disclosed.

“Today’s announcement is another step in Siemens’ commitment to help our customers fully realize the benefits of digitalization,” said Chuck Grindstaff, CEO and president of Siemens PLM Software (Plano, TX), in a statement. “By adding Polarion ALM solutions to our PLM portfolio, we are further strengthening our ability to help companies create smart, connected products.”

New Releases

Rockwell Automation (Milwaukee) Nov. 5 announced the addition of applications for its Rockwell Software Studio 5000 environment to help engineers speed development of automation systems as they design a Connected Enterprise. These applications, along with the Studio 5000 Logix Designer software released in 2012, bring more functionality together into one environment to help improve automation design productivity.

Rockwell’s latest additions include the Studio 5000 Architect application, where users can view the overall automation system; the Studio 5000 Logix Designer application, the design and maintenance software for the Allen-Bradley Logix5000 family of controllers; the Studio 5000 View Designer application for Allen-Bradley PanelView 5500 graphic terminals; and the new Application Code Manager that speeds system development in helping users build libraries of reusable code that can be deployed across the enterprise.

RoboVent (Sterling Heights, MI) announced Nov. 9 the release of eTell, an advanced predictive control system for air-quality equipment that improves energy efficiency, prolongs filter life and reduces operational costs. RoboVent, a builder of industrial air-filtration systems, called the cloud-based control program the first of its kind in the industrial filtration industry, using advanced analytics to “learn” the plant’s processes and trends and make automated adjustments to improve system performance thus increasing efficiencies and lowering costs.

With the eTell system, plant managers can easily monitor energy use, maintenance requirements and system usage with the eTell mobile app and dashboard. “eTell moves beyond preventative maintenance and in to predictive maintenance,” said Jim Reid, RoboVent general manager. “With eTell, plant managers don’t have to wonder when maintenance is needed. The equipment can reach out and tell them when it needs something, whether it’s a new filter or a maintenance check.”

eTell connects to RoboVent equipment via Bluetooth and its smart software continually learns from system or human input and makes predictions and adjustments based on that input, similar to Google Nest thermostats. The new system is the latest in a series of systems that provide greater visibility, control and efficiency to plant managers, building on successes of other electronic control systems from RoboVent such as eDrive, an automated Variable Frequency Drive (VFD) program that adjusts power based on filter load, and eMaster, a facility-wide electronic monitoring system.

  • View All Articles
  • Connect With Us

Related Articles

Always Stay Informed

Receive the latest manufacturing news and technical information by subscribing to our monthly and quarterly magazines, weekly and monthly eNewsletters, and podcast channel.