( This article is excerpted from the complimentary report: IoT: From Hype to Adoption — Five Ways Manufacturers and Distributors Are Adopting the Internet of Things, available for download here.)
IoT — From Hype to Adoption
The Internet of Things (IoT) is moving out of the hype cycle into early mainstream adoption. Many emerging and growing manufacturers and wholesale distributors are working to determine where IoT fits into their business strategies. A key part of this is prioritizing the numerous potential IoT initiatives. Many of these companies face constraints in budgets, inhouse expertise, bandwidth, and technical resources. This makes it even more critical for a growing company to choose wisely on where to start and where to invest in IoT. We look at five major areas that manufacturers and distributors are implementing IoT: 1) on the plant floor, 2) in supply chain and logistics, 3) in service and repair, 4) incorporated into products, 5) creating value-added services.
1) On the Plant Floor
Manufacturing plants have been instrumented with sensors for many decades. In the past, those sensors were often only used for local control of a machine and process. For example, CNC1 machines have been in use since the late 1950s. Process control mechanisms have been around even longer (for centuries) but have seen tremendous advancements in the last 50 years. The systems and applications that control factories are often referred to as Operational Technology (OT). These systems tended to be proprietary, isolated, and hierarchical in nature. The Purdue Enterprise Reference Architecture (PERA) defines a five-level hierarchy for plant and production control related technology.
IoT brings greater connectivity, peer-to-peer/machine-to-machine communications capabilities, more standardization, increased analytic capabilities, and greater connectivity to and integration with enterprise systems. This last phenomenon is often referred to as IT-OT convergence. Thus, IoT can be seen as an evolution building on many IoT-like capabilities already in the factory. IoT is primarily augmenting, rather than replacing the systems in a factory. For example, SCADA technology vendors have been adding IoT capabilities to their systems which allows for increased automation and control within the factory.
A High Bar for Cyber Security on the Plant Floor
IoT systems typically have one or more IoT gateways, which translate proprietary sensors’ outputs and industrial system protocols into standardized protocols and data. This makes previously siloed data more broadly accessible. Many of these previously siloed systems were never designed to be connected to the network and hence lack the rigorous security mechanisms required to survive on today’s internet. Robust, multi-level security mechanisms have become imperative,2 especially for factory systems where the consequences of breaches and sabotage can be detrimental.
Production Quality, Analytics, and Autonomation
The availability of additional data in real-time provides increased ability to generate operational alerts. The capability to analyze previously non-integrated data enables ongoing process improvements. IoT can be used to reduce variability in manufacturing processes, ensuring proper operations and equipment settings, making automated real-time adjustments, and verifying that procedures are faithfully adhered to. Vision-based inspection systems and precise automated measurement systems are used to improve quality. IoT enables higher levels of factory automation by providing sensor data and connectivity to guide machines and robots.
IoT can provide visibility into what is happening across a network of factories, including your own factories, outsourced manufacturers, and suppliers’ factories. For example, some apparel vendors use RFID to track, in near-real-time, production status and events at their supplier’s factories.
2) In Supply Chain and Logistics
We are starting to see higher adoption of sensors and locating technologies in warehouses and distribution centers, transportation, and at supplier’s factories.
IoT in the Warehouse
A modern distribution center (DC) often already has numerous sensor-based technologies. While some don’t consider RFID and wireless connected barcode scanners to be IoT, in fact they are connected sensors that provide granular real-time data about the locations of items, cases, pallets, bins, vehicles, and workers in the factory — a digital twin of the operations. The increase in ecommerce and small order fulfillment has strained the ability to find enough warehouse labor, leading to increasing automation. Amazon’s acquisition of Kiva (robots that bring the items to the order picker) has accelerated investments in order picking automation (or semi-automation). Systems from other solution providers, such as Locus Robotics, provide Amazon’s competitors with similar robot-aided picking capabilities. Humans are still needed to pick the items, but these systems have increased productivity by 3X or more. In general, IoT devices also help dramatically reduce picking errors. For example, a smart connected scale can be used to weigh the picked order before the worker seals the box for shipment. Information about which items are being shipped, how much each weigh, and how much the associated packaging should weigh, is used to ensure the right items are in the box.
Tracking Products Through the Supply Chain
The traditional approach to tracking the status of shipments through the supply chain involves a combination of EDI messages, emails, and phone calls, with accompanying delays and inaccuracies in the information. Major parcel carriers use IoT technologies (GPS, connected scanners) to more precisely track the status of packages. IoT offers the potential for more real-time tracking of products throughout the supply chain, but not without challenges. For one thing, the trucking industry is highly fragmented — there are over 37,000 trucking companies in the American Trucking Association alone, and the US Department of Transportation has over half a million for-hire carriers (trucking companies) on file. Many of these are mom and pop shops with no tracking capabilities. This makes end-to-end tracking challenging. Several networks3 have emerged that leverage drivers’ phones and/or in-vehicle devices to track the location of vehicles.
End-to-end tracking, especially for overseas shipments, can be even more challenging, as it involves multiple handoffs. There are several startups4 that are trying to address this by attaching a GPS device to containers or pallets to track them from source to destination. Different companies use different backhaul technologies to send the location information, either continuously or at waypoints. These include cellular, satellite (expensive), AIS (for the ocean leg), WiFi, and even Bluetooth. These are often combined with GIS5 systems, such as from esri.
Condition Monitoring and Perishables Supply Chains
Temperature, humidity, shock and vibration, and other sensors can be used to monitor the condition of goods travelling through the supply chain. These are typically incorporated into service offerings from carriers, 3PLs, or solution providers. For perishable food items, companies like Zest Labs offer not just end-to-end temperature tracking, but the shelf-life models, process modeling, and intelligent distribution algorithms required to squeeze waste out of the system. Typically, DCs use FIFO (First In, First Out) algorithms to ensure stock rotation. More sophisticated operations use FEFO (First Expired, First Out). With accurate shelf-life models, based on end-to-end temperature tracking, the true number of remaining days of freshness is known. Temperature tracking can also be valuable for temperature-sensitive pharmaceuticals, such as biologics. In that case, it becomes critical to know the extent of any temperature excursions, as lives may depend on efficacy of the drugs and degradation cannot be observed from visual inspection alone.
3) In Service and Repair
Predictive maintenance is enabled by IoT sensors embedded into products in the field and equipment in plants, combined with machine learning algorithms that find predictive patterns in the sensor data. Traditionally, equipment is repaired based on elapsed time or hours of use. A margin of safety is built into the repair period guidelines. As a result, many parts are replaced, and repair procedures are performed long before needed, while other machines fail before being serviced. With predictive maintenance, machine learning algorithms look at various sensor data (temperature, vibration, particles in lubricants, etc.) to accurately predict when maintenance is needed, in line with actual wear and tear. This results in fewer scheduled maintenance calls, using less labor and inventory, while simultaneously reducing failures. It reduces the number of emergency situations that require very expensive logistics and labor.
According to a DOE survey, implementing a functional predictive maintenance program can reduce equipment breakdowns by 70%-75%. Predictive maintenance and IoT embedded into products provide a foundation for servitization, moving from selling products towards selling outcomes.6 This topic is covered further in Part Two of this series, where we discuss how companies are incorporating IoT into their products in order to sell product-as-a-service and provide layers of value-added services on top.
1 CNC = computer numerical control systems, which use synchro (aka Selsyn) mechanisms to sense and precisely control the position of cutting heads on machining equipment. — Return to article text above
2 For more on this topic, see The IoT Security Imperative. — Return to article text above
3 Such as MacroPoint (acquired by Descartes) and FourKites. — Return to article text above
4 Such as Tive and ODYN — Return to article text above
5 GIS = Geographic Information Systems — Return to article text above
6 For more on outcome models, albeit with a software slant, see Outcome-based Business Models for Enterprise Software — Return to article text above
To view other articles from this issue of the brief, click here.