Subscribe and listen to AW’s podcast!
Subscribe and listen to the Automation World Gets Your Questions Answered podcast!
Listen Here

Taming the Data Monster

With the right strategy and tools, you can master your data and put it to work improving business performance.

Aw 100318 Datagraph

Factories today are smarter than ever before. Intelligent motors, smart drives, safety PLCs, sensor networks¾the connected devices in the industrial Internet of Things are on track to reach 48.2 billion installed by 2025. Much of the value of these devices lies in their ability to communicate, delivering frequent updates about their status, performance, and availability. They generate vast amounts of data, the majority of it a variations on "I'm fine, I'm fine, I'm truly fine, I'm very fine, I’m oh so fine, thanks for asking.” As with those 1950s creature features with ants the size of pickup trucks, something easily managed when it's small can be downright monstrous when it gets too big. That onslaught of data from the production floor can attack your budget, piling up charges for engineering hours, storage hardware, energy to run that hardware, space for the hardware, and network bandwidth.

The sheer volume also introduces the risk that all of those messages (“I’m still fine”) will overshadow an anomalous reading that might otherwise allow you to catch a fault before it happens. Fortunately, a careful data strategy coupled with visualization tools can help you tame the data monster and convert all of those bits and bytes into actionable insights.

Have a Plan

The first step in taming the data monster is to avoid creating the monster in the first place. Although it can be tempting to monitor everything you can in the spirit of keeping your options open, that approach is counterproductive. It’s better to begin with a single issue, identify the data necessary to address it, and craft a solution. Maybe the goal is to track energy consumption for each machine or to compare operational equipment effectiveness (OEE) from facility to facility. By executing your initial project, you and your team will learn the process. You’ll develop the skill set necessary to attack future problems, because solving one issue inevitably leads to identifying new ones.

Visualization applications are aimed at distilling down masses of data and making it consumable by the greatest number of job roles throughout the organization. To fulfill that task, first and foremost it needs to capture input in quasi-real time from assets across the enterprise. In addition to manufacturing equipment, the system should be able to monitor support systems like cooling towers, chillers, boilers, and air-handling units.

The goal is to uncover hardware, production, and supply-chain trends that can be used to optimize performance at all levels. Fault-detection software, for example, leverages fault-detection algorithms and preset thresholds to more efficiently identify unusual behavior and potentially problematic trends. It might reveal that the machine has a new resonant frequency that indicates a mechanical problem in the making or that a drive is drawing progressively higher current and needs to be checked. Timely access to that type of information can enable quick action to prevent line stoppage.

Fast access to insights

The true value of visualization lies in presenting key performance indicators in a way that enables stakeholders to easily access it in context and take action. Maintenance might need to see faults, error codes and spares inventory on a tablet, for example, whereas a facilities manager may be more interested in viewing energy consumption and throughput on their laptop. Applications built for specific analysis and reporting purposes like quality, energy, and fault management are designed to take very large amounts of data and put them into formats that are easily digestible and actionable. Alternatively, developer tools are designed to simplify and speed the process of building home-grown versions.

Modern graphics packages can be used to develop rich, immersive 3-D renderings. This isn’t just restricted to data plots. These suites can create detailed models of the assets on the factory floor, linked to status information and performance data. Users can scan and zoom to view different equipment, clicking to drill down into quantitative information. More sophisticated versions combine the plant model with satellite mapping data to enable management to monitor widely-separated assets at a glance, pinpointing the location of faults in seconds.

Consider alarm management for a global food and beverage company. Imagine that you could view a world map that marks each facility with a colored dot (green, yellow, red) to show the operating status. When you see a red dot marking the Peoria plant, you click on it to access the 3-D model, which shows that the bottling line is shut down. By zooming in, you can see that one of the doors on the machine is ajar. The software doesn’t just streamline diagnostics, it enables you to see how long the line has been down and how much production has been lost. You can also check the status of other plants to see whether they can make up for the shortfall.

The most effective way to deliver this information is on customized dashboards that present multiple metrics on a single display. When tailored to the job role – machine operator, plant manager, maintenance tech, system integrator – these support rapid decision-making. To take the above example one step further, the fault detection information could appear in just one corner of a web-based visualization environment. The top right corner might display energy consumption, while the bottom left corner could show overall equipment effectiveness (OEE) and the bottom right corner provide access to historical trends for any of a number of key performance indicators. Anytime you want more detail, you can drill down to see quantitative data by enterprise, facility, machine, and even shift or operator.

Data is valuable only to the degree that it can be converted into information. Today’s intelligent factory can generate a frightening amount of data. If not managed properly, it can not only fail to benefit the enterprise, it can cause active harm. Using the latest analysis and visualization techniques, end users and OEMs can conquer the data monster, converting the data stream into insights to improve performance of the organization, top to bottom.


Further Reading



Previous issues of Creating Productivity

Making the factory of the future, today

Cutting Total Cost of Ownership


Meet the tools:

AX Energy

Conserving energy is the right thing to do, both environmentally and economically. Given the amount of data and the complexity of the analytics involved, however, it can be challenging to pull off. Enter Mitsubishi Electric’s AX Energy, a web-based energy-management application that can slash utility bills by up to 25% annually. Designed for easy deployment and fast ROI, it analyzes the output from thousands of meters to present key performance indicators in a rich visualization environment. The Site Summary Overview, for example, presents details of overall energy consumption, cost, and carbon footprint. The ISA95-compliant Asset Tree feature makes it easy to drill down to reports and charts for equipment and locations throughout the operation. Because the application is web deployed, users can access it from a range of platforms, including PCs, mobile phones, tablets, and HMIs. It’s available out of the box with preset calculations that can be easily customize to suit specialized requirements

AX Facility

Mitsubishi Electric’s AX Facility provides an easy-to-use application to convert data into actionable insights. The visualization software is based around real-time fault detection and diagnostics (FDD) algorithms that monitor equipment performance and operation and convert it into easily consumable output. AX Facility calculates the probability of fault occurrence and the associated cost, making it easy to predict downtime and prioritize the correction of faults based on available resources. The application includes a library of more than 300 preconfigured fault rules for common manufacturing and facilities automation equipment. Customization options allow users to modify these rules for their particular operation to get up and running quickly.

Fault-state tracking provides traceability so that maintenance and managers can easily access the frequency, duration and resolution of faults. Automated fault detection delivers real-time notifications to even remote locations and the FDD Viewer option presents list of probable causes. AX Facility has been shown to reduce service costs by 50% and energy costs by 6% or more. Users typically see return on investment in 12 to 18 months.

AX Portal

The true value of factory visualization lies in the ability for users to view and assess multiple metrics at the same time. AX Portal is designed to deliver on that process. It enables the rapid development of customized dashboards aggregating input from visualization applications such as AX Energy, AX Facility, and modules from MC Works64. It can also interface with common enterprise databases like SQL, Oracle, and SAP to perform analytic calculations in its own right. A library of .Net automation Web parts has been developed with sections targeting visualization, data access and operations, and reporting. The resultant displays go beyond just numbers to include maps that can be drilled down to 3-D models of assets while data plots present the associated metrics. Want to move things around? Drag-and-drop functionality makes it easy. Now you have the tools to efficiently assemble a unique dashboard tailored to deliver secure, role-based data to every group of users in your organization.

Test Your Machine Learning Smarts
Take Automation World's machine learning quiz to prove your knowledge!
Take Quiz
Test Your Machine Learning Smarts
Discover New Content
Access Automation World's free educational content library!
Unlock Learning Here
Discover New Content