Leveraging Advanced Analytics to Create Smart Factories

Leveraging Advanced Analytics to Create Smart Factories

Reading time: 6 mins

by Kumar Singh, Research Director, Data & Analytics, SAPinsider


Harnessing the true value of Industry 4.0

The promises that Industry 4.0 foundational technologies bring to the world of supply chain and manufacturing are many and are attractive. Smart factories and smart manufacturing processes will be an integral component of almost all Industry 4.0 networks. Embedded seamlessly in an Industry 4.0 network, these smart factories will not only help realize optimal, flexible, and agile manufacturing within new deployments, they will also share important data points with the overall network. This seamless exchange of information will ensure that the entire network leverages the full value of having Industry 4.0 capabilities in place.

The journey to smart manufacturing revolution

Building smart manufacturing capability is not easy though. It is a journey that needs to be carefully planned and orchestrated across a diverse set of data inputs like people, processes and technology. Each of these elements need to be in complete sync with each other to realize the vision of a true smart factory. As mentioned previously, a true smart factory will not only completely transform the manufacturing paradigms of an organization but also transform related functions like marketing, engineering, and operations.

One very critical aspect of starting your journey towards building smart factory capability is to understand what exactly that capability means. Often, manufacturing automation and smart manufacturing are used interchangeably by organizations when highlighting these capabilities. However, it is critical to understand that automation, while it may lay the foundation, is not smart manufacturing. Then there is another notion that smart manufacturing is about getting near real time visibility of your manufacturing operations. Again, while this may be an essential component of developing smart manufacturing capability, on its own, it is not an indicator of smart manufacturing capabilities. Smart factories are combination of many aspects like automation, near real time visibility, and most important of them all, inferring something useful out of that data,  aka analytics. SAPinsider recently had the opportunity to sit with Intel’s Industrial Edge Insights Director, Bridget Martin, to discuss the key aspects of smart manufacturing platforms and how data science plays a key role in developing smart manufacturing capabilities.

The real value is in the data

The value of insights generated from data is not new in the world of manufacturing. Best of breed companies have been leveraging analytics on data generated and captured by their manufacturing operations to minimize manufacturing related risks, improve manufacturing efficiency, reduce manufacturing waste, improve product quality, and at a strategic level, make better business decisions. However, a key aspect of these analytics approaches in the past was that they were performed on latent data and hence, many actions that companies took based on the insights generated, were reactive. As businesses evolve rapidly in today’s digital age, organizations can no longer afford latency. Consider the example of product quality. If a product quality issue goes undetected and the product gets in the hands of customer before the data driven insights identify the issues, the impact on brand and customer loyalty can be huge. In the “Amazon age”, customers are extremely demanding, and products are getting commoditized so fast that it may not take much for your customers to switch to your competitors. And this is where the need for near real time monitoring and analytics of data generated by manufacturing processes comes into prominence. What organizations need are solutions that can not only help them get near time visibility into their data, but also be able to leverage the data to transform processes and business models.

As Bridget Martin, states: “There is a need for ready-to-deploy reference software platforms designed for near realtime, minimal latency video and time series data analytics that enable factory owners to automate and advance their operations—without replacing existing machinery, production lines or processes (“plug and perform”). Factory upgrades can be delivered via software updates that factory owners can manage remotely and deploy automatically, potentially reducing costs, saving time, and expanding capabilities more easily.” Organizations realize this need, and so do the solution providers. More and more platforms are emerging in the market that promise to help companies realize their smart manufacturing vision. But one aspect that is obvious is that not all of them are the same in terms of capabilities. And the most significant capability that differentiates these solutions is the “smartness” that the solution contains, which is in the form of algorithms built into these solutions.

Smart manufacturing platforms fueled by algorithms

As mentioned above, an ideal platform should not only aid in automation but should help take the capabilities significantly further, by embedding algorithms in the platform that organizations can use to gain insights, optimize processes, and even develop new capabilities, like smart quality management programs. These algorithms can range from conventional optimization algorithms to advanced deep learning algorithms. Since these platforms cover end to end manufacturing processes, the portfolio of algorithms available in these platforms need to be diverse as well.

Bridget highlighted this in her quote: “The ideal software reference design should include sample algorithms for various use cases or easily plug in third-party/open source developed algorithms, and even enable customers to develop their own algorithms if needed with its built-in training and learning tools. Designed exclusively for manufacturing environments, the platform should have unlimited manufacturing use cases—whether discrete process applications like electronics or auto manufacturing or process automation applications in the Oil and Gas sector.”

Interoperability adds to the complexity of Industry 4.0 world. To develop a true Industry 4.0 network, platforms need to “talk” to each other and hence this is a feature that is very critical. Bridget quoted some example features like: “Capability to push and publish AI analysis to local applications or the cloud. Should have containerized microservices which are easy to modify and customize for a factory owner’s unique applications. It should be easily adapted, extended, and scaled across operations.”

Transformation example: smart quality management at Audi

During the discussion, Bridget highlighted an example of the transformative capabilities of a smart manufacturing platform. The example was around how Audi was able to leverage a smart manufacturing platform to develop a smart quality management process. “At Audi, an individual car has a significantly large number of welds. Using a smart manufacturing platform, Audi was able to automate and expand its quality inspection processes to inspect 100% of welds in the factory and more efficiently, with an estimated 30-50% immediate reduction in labor costs. For welds outside the quality guardrails, Audi can easily tell where they are in the factory and act more quickly to address them. “

What does this mean for SAPinsiders?

The road to building smart factories is not straightforward. Careful planning and strategic selection of tools and external partners will be critical to developing this capability which most manufacturing organizations are aiming to build to successfully compete in the digital age. Some aspects that you need to be cognizant of are:

  • Evaluate your foundations. At the core of smart manufacturing capability are four key ingredients-Connectivity, automation, visibility, and analytics. These four key ingredients must come together, building the foundation so it is imperative that you evaluate your current state in these areas, to understand where you stand, what is the delta and what needs to be done to cover that delta.
  • Invest in your people and business processes. People and processes are as important as technology in any initiative and more so when you want to build any new capability enabled by technology. While it may not always be required, sometimes you may need to redesign your processes to ensure that you will be able to leverage the full value of any smart manufacturing platform. Human machine collaboration is going to be a critical aspect of Industry 4.0 networks, so you need to ensure that your manufacturing talent has the skills to leverage the fill value from these platforms.
  • Take time to evaluate solutions. The smart manufacturing platform that you leverage will be one of the central components of your smart manufacturing capability to you must put together a comprehensive evaluation criterion in order to select an optimal solution.


Kumar Singh, Research Director, Data & Analytics, SAPinsider, can be reached at kumar.singh@wispubs.com

About Intel® Edge Insights for Industrial (Intel® EII)

Intel® Edge Insights for Industrial (Intel® EII) is a pre-validated, ready-to-deploy reference software designed for near realtime, minimal latency video and time series data analytics that enables factory owners to automate and advance their operations—without replacing existing machinery, production lines or processes. For more information, visit Intel Edge Insights for Industrial

More Resources

See All Related Content