Methodically Improve Your Factory Floor Analytics

Picture a complete digital representation of your factory. You're controlling and monitoring everything from robotics to room temperature while running what-if scenarios on computers that look like they came straight out of NASA's control center.

In a perfect scenario, the entire factory ecosystem communicates, from factory floor machines alerting machine engineers of potential problems, to algorithms telling the machines what to do next.

The immediate reaction for some manufacturers is to believe this type of smart factory is out of reach. That's not the case at all. If your organization starts taking smaller steps and builds on them, over time, you’ll be sitting in front of that control center.

Your manufacturing organization should modernize to remain competitive. The returns that data analysis, automation, optimization, and monitoring bring to factory operations are worth the investment.

If you can analyze your factory operations and improve the yield, your margins will go up. Your operational expenses will decrease if you leverage artificial intelligence (AI) and machine learning (ML) to automate tasks. And if you can increase throughput and improve customer lead times, revenue will increase.

The previous article in this series, Three Steps to a Better Overall Equipment Effectiveness (OEE) Program, outlines steps designed to be implemented on one machine and replicated across all line equipment to improve production throughput. It's a great place to start.

In this article, we move on to automating the data capture for systems in the greater manufacturing ecosystem, such as ERP, MES, and CMMS. Combining this data with machine data begins the path to greater production optimization—with factory dashboards and actionable analytics.

Factory floor systems and analytics—deciding what data to capture first

Auto-capturing data across factory floor systems, both machine and process related, creates new opportunities to achieve high levels of productivity and specialization. Fast, efficient data analysis becomes possible.

However, a typical factory floor has many systems and lots of data. You want to avoid a project that tries to capture all the data from all your systems simultaneously.

The key is to figure out what application to start with and what data to capture. As you'll notice, I don't use the term “big data” because the concept is to take a targeted data approach, so you avoid the paralysis that comes with large amounts of data sets.

Here’s a high-level overview of the steps that have been successful for our clients.

To turn your smart factory goals into a reality, begin with a discovery and needs analysis, mapping out your systems and critical data. Next, bring in the business subject matter experts to help develop the data strategy roadmap.

Finally, have your team evaluate the roadmap and prioritize each step based on its potential business impact. If done right, it will be clear what data should be captured first and from what system.

Let’s say your team decides that machine breakdowns are the biggest problem, then combining machine data with your CMMS is the logical next step. And if the next big problem identified is product production numbers, then analyzing the correct data from the ERP system with machine production data might be next in the roadmap.

The data strategy roadmap will also identify operational bottlenecks and potential savings opportunities across the factory and its production lines. Once the roadmap process is completed, you're ready to build a logical project plan.

Our customer, a leading global discrete manufacturing company, found this methodology to be a resounding success, and that’s not unique. It’s how we work with our customers to move them steadily toward their smart factory goals. Here's a synopsis of their success story.

The demand for this customer's product was increasing quickly. In response, they needed to increase product output without adding additional factory space. They knew it could be done using factory data analysis to find bottlenecks and streamline production processes.

They followed our methodology of starting with automated machine data capture to substantially improve overall equipment efficiency (OEE) while creating a roadmap for future returns.

SoftServe is leveraging Google's MDE platform to collect and manage data, then applying AI and data analytics to better understand the client's production processes, and finally, implementing algorithms to optimize and deploy using edge computing.

The project timeline steps consist of three stages:

• Automate data capture from their factory machines
• Use the data collected for machine learning model training
• Analyze the data acquired to optimize specific machine operating rates

The roadmap we provided this client will modernize operations, giving them:

• The ability to connect their multisite factory wholistically
• Real-time visibility into their complex product assembly process
• AI and ML to improve inspection capabilities across the factory and assembly process

With this foundation, they increased product output and can now focus on moving toward their long-term business intelligence goals, with a path to their vision of a smart factory.

Building a smart factory is within reach. Ask a SoftServe expert how you can get started on a roadmap to success.