Industry Week says the IoT is a game-changer because it provides component-level over batch-level visibility. That’s one use for the IoT on the factory floor; the other is to gather big data and run analytics to improve planning, operations, and supply chain efficiency.
Here’s how manufacturers can use the IoT to do both, and why they’re so important for factory optimization.
Factory Optimization: The Long View
Optimizing factory processes is, by nature, a mathematical problem. It fits perfectly into the IoT and big data analytics models, which are designed to gather quantifiable metrics to provide insight and protection.
Mathematicians would probably claim that setting the optimal levels to run factories is a linear programming problem. Linear programming determines the optimal levels of inputs and outputs given certain constraints, and describes the assembly line and production planning, both of which are governed by capacity constraints, speed and the timing of component availability.
Since early motion studies, many have tackled the problem of optimizing factories, the most famous being W. Edwards Deming. Douglas MacArthur sent this unknown American statistician to Japan to rebuild their industry after WWII. He literally taught Toyota how to build cars; manufacturers from around the world now study “The Toyota Way.”
The next group to examine how to fine-tune factory operations and assembly lines consisted of operation researchers, but their impact was not significant, and their correlation, classification, and machine-learning models were too complicated to understand.
Enter Big Data and IoT
But all of that has changed thanks to the IoT, big data, and analytics. All three have made it possible to use applied mathematics on factory floors without needing to understand the logic behind respective algorithms.
Over the past few years, databases like Apache Spark have emerged to help businesses process streaming data in a way old-fashioned, rigid data-warehouses could not, and new programming SDKs have made it possible for ordinary programmers to apply operations research data models and machine-learning algorithms to the problems of optimization. IoT clouds and SDKs transmit data to these databases and algorithms through network-connect sensors attached to components and machines.
Tracking the Assembly Line
There are an infinite number of ways to tap into the real-time flow of information coming from factory floors, warehouses, and supply chains, yet manufacturers have been somewhat slow to embrace the digital model. This leaves serious room for improvement, and when manufacturers decide to move in this direction, the first items they’ll likely focus on are preventive maintenance and logistics given how simple they are to optimize.
General Electric is one company that has embraced the IoT. Not only does GE have 10,000 IoT sensors at their Duracell plant in Schenectady, N.Y., they place a sensor into every battery.
IoT-equipped batteries radio their location as they move across factory floors and supply chains. This renders manual barcode scanners obsolete. Additionally, sensors measure weight, volume, ambient temperature, vibrations, noise, pollution, and light levels in machinery and components throughout the battery manufacturing process to ensure operations remain on track. Variations in any of these factors are usually a sign that an operation is out of bounds.
The Simplicity of Going Digital
IoT vendors, analytics cloud companies and articles in Forbes are quick to say the IoT and analytics will “speed time-to-market, reduce TCO, and improve customer relations,” but what often goes missing from vendor landing pages and articles are precise steps on how this will occur.
The technology and tools have become cheap and easy to deploy and understand. Sensors themselves only cost about $1. An Intel Edison computing card costs approximately $35, and evident simplicity and low costs remove any need for network engineers to bring them online. ZigBee, Z-Wave, and BLE (Bluetooth 4) transmission protocols can work in peer-to-peer fashion. Thus, only one end of the network needs an IP address or controller.
Some companies have even begun attaching sensors to humans to track the efficiency of their motions and where they linger. Of course, not every employee would favor that.
Then there’s the IoT cloud, which manages IoT devices. It produces software updates and instructions, and allows operators to reboot as needed.
The Planner’s Dashboard
Data gathered in this manner from factory floors, warehouses and supply chains flows into easy-to-read dashboards. This allows planners, managers and assembly line workers to visually track what is happening and make changes accordingly. That approach is miles better than viewing metrics gathered after the fact, which is exactly what companies did before the IoT and big data. This modern approach eliminates the lag time between finding problems and appropriately fixing them.
But digitizing the factory floor is only half of the equation. The company-wide analytics model adds transactional data coming from the ERP and CRM systems, consumer buying-pattern data purchased through data brokers, and brand perceptions on Twitter. This data then feeds into predictive algorithms and machine-learning models. This paves the way for tactical and strategic planners to react to changes accordingly.