How the internet of things is changing manufacturing
It is estimated that by 2020 there will be over 20 billion connected devices. By 2030, that number could grow to as many as 500 billion. Industrial manufacturing is expected to account for as much as 62% of IoT technology deployment. It is changing manufacturing today—and will continue to do so at an ever-increasing pace. Many manufacturers have started their journey into IoT and have begun to realize true value for their operations. The financial return on investment is certainly significant, but “soft cost” savings such as human efficiency should also be understood and publicized. The key to a successful IoT deployment is the people implementing and using the tools. By using IoT technology, manufacturing operations can improve their productivity and make their employees’ jobs easier.
Key areas of impact within manufacturing
Data availability and utilization
As the number of devices gathering information goes up, so does the amount of data available. This data can be used to rapidly make smart decisions on the manufacturing floor. Certainly, the data displayed should be restricted to only the most useful insights. Smart dashboarding can help the manufacturing team meet its goals, by enabling adjustments throughout the day based on quantitative performance data.
Once the collected data is sent to the cloud, it becomes available throughout the organization. This enables offsite personnel and management to view data while it’s fresh. It can also save the local group time, as they no longer need to quickly drop everything and generate a report for management.
Once the data is robust, it can be leveraged for advanced analytics. This can provide deep insights into the manufacturing process. One of the methods to perform this analysis is through machine learning.
Many decisions can be made autonomously by the machines themselves to optimize production. These are typically decisions that are difficult to make by humans. Rather than sifting through mountains of data in spreadsheets, we can use machine learning techniques to find the patterns of importance.
The power of machine learning is that it relies on empirical data to draw conclusions. The factory’s engineers don’t need to derive equations and apply theories; they can use real plant conditions to feed the algorithms.
Higher precision in lot monitoring and raw materials can establish the optimal production rates without over- or under-producing. The improved efficiency rate can be honed and improved in real time without human intervention.
Data analysis can also yield optimal changeover patterns and steer the deployment of human assets, reducing idle time throughout the factory. Harley-Davidson used these processes to reduce build-to-order cycles by a factor of 36 with a fully IoT-enabled plant.
IoT technology also enables better traceability for raw materials within the supply chain. In one recent case, Walmart incorporated IoT with blockchain technology to require manufacturers of food products to supply complete traceability from the retail floor back to the grower. This enables food retailers to know more quickly which lots must be recalled.
Drone technology allows for inspection and mapping in hard-to-reach areas. Companies like Duke Energy are using drones in several ways: inspections on remote and widespread equipment, thermal imaging and construction tracking. By using drones, the company receives more accurate inspection records faster. They take employees out of potentially dangerous areas, reducing worker risk. There are also companies using drones to build accurate and up-to-date three-dimensional models of their facilities. These models can be used to track construction progress, for training or to troubleshoot an issue.
The term “digital twin” has many meanings, and it can sometimes be confusing. One potential example builds on our previous discussion on drones. Once a 3D model of a facility has been built, it can be used for troubleshooting. Imagine a scenario where a steam trap fails in a refinery. There are potentially thousands of steam traps at the site. Which steam trap was it, and where should the maintenance technician go? What if an outside vendor needs to be involved? This can now be done remotely using the model to quickly locate the position.
Another very different use case for the digital twin is in simulation. In this instance, the digital twin is a simulation of the process. It can then be used to train operators on real scenarios so that they become comfortable reacting to unusual conditions. It can also be potentially used to find opportunities to improve the process. Operators and engineers can try out scenarios and determine whether they constitute an improvement.
Smart condition monitoring
IoT technology is also enabling manufacturers to move toward condition-based predictive maintenance and away from less efficient preventive maintenance schemes. With older, time-based maintenance programs, service was based on statistical analysis often provided by the OEM for each piece of equipment. With predictive maintenance, IoT-enabled equipment can use real-time data to determine in advance what maintenance needs to be performed to reduce downtime. One study by Deloitte found that predictive programs can reduce breakdowns by 75% and reduce maintenance costs by 25% while improving overall factory productivity by 25%.
Central to these predictive maintenance programs are smart condition monitors (SCMs). In a complex manufacturing environment, there are often hundreds, even thousands of mechanical components and rotating equipment. This equipment has traditionally relied upon vibration sensors and other monitors to detect mechanical failure or variance. However, using modern smart condition monitors, sensors are now capable of monitoring assets from within the motor control cabinet. By detecting faults in AC induction motors and rotating assets at an early stage, SCMs enable maintenance strategies to become proactive, saving technician labor time and money—and raising valuable uptime.
High-performance SCMs are also plug-and-play, utilizing cloud-based analytics and intuitive dashboard readouts to provide complete maintenance visibility. Instead of being deployed near the asset, SCMs are installed in the motor control cabinet, meaning they can be deployed at a large scale on assets in harsh conditions and still deliver the required actionable data. The result is less unplanned downtime, optimized performance of rotating assets, and reduced energy waste.
Mitsubishi began using a smart condition monitoring system in 2017 and operators are now able to monitor rotating equipment status throughout the factory. The system sends alarms on a predictive basis ahead of impending failure and sends data for recommended actions.
Continued IoT growth
As IoT technology scales across manufacturing, its impact has already been realized. At the shop floor level, it enables companies to gather data on labor utilization, equipment utilization and quality through connected devices that operate at a rate and accuracy not possible for human analysis. Combined with predictive analytics, these devices are making a true “smart factory” possible in which humans, machines and algorithms empower each other to achieve more.
Semiotic Labs’ SAM4 is a smart condition monitoring system for AC induction motors and rotating assets. It is an end-to-end solution that includes sensors, analytics and an online dashboard (or API) that offers clear, actionable information about the health, performance and energy consumption of connected assets. For more information, request a SAM4 demo.
Resources58 See all resources
Semiotic Labs Named A Cool Vendor in Gartner’s May 2020 Cool Vendors in Manufacturing Industry Solutions Report
Semiotic Labs, a leading provider of predictive analytics to eliminate industrial downtime, today announced that it has been named a Cool Vendor in Gartner’s May 2020 Cool Vendors in Manufacturing Industry Solutions report.
Condition monitoring of critical pulp and paper production processes can alert the maintenance team to a developing fault before that fault causes unplanned downtime.