About the author:
This post was guest-authored by Kayla Matthews. Kayla is a writer from Pittsburgh, PA, who has contributed to sites like IoT Magazine, Born2Invest, Ubidots, and more. To see more future IoT and business articles from Kayla, visit Productivity Bytes or follow her on Twitter and LinkedIn.
Technology has helped kick off what many are calling Industry 4.0. Machine learning and the Industrial Internet of Things (IIoT) both play influential roles by powering robotics, automation, enterprise resource planning, and cloud computing and collaboration software. However, this high-tech infrastructure doesn’t have to replace humans — it can complement and assist them.
The IIoT gives manufacturers “eyes” and “ears,” to gather data, the cloud makes that data mobile, and machine learning provides tools for interpreting the data and making predictions. Here’s a look at how, and some of the ways industrial manufacturing stands to benefit.
1. Detecting and Correcting Errors
Not long ago, quality control was an extremely labor-intensive process. However, operations became more sophisticated over time, with machinery able to spot their own errors. However, there’s still a chance for defects to slip through if employees don’t follow inspection protocols.
How much longer will human beings carry out this kind of work? In some tests, machine learning has shown it can improve error detection by 90 percent compared to people performing the same task.
At face value, this doesn’t sound like a positive thing for the human workforce. However, multiple factories have shown that automating vital functions, like defect defection, doesn’t have to come at the expense of jobs.
Machine learning and IIoT-based cloud computing come together nicely here. Machine learning and the IIoT provide the “self-awareness” by which machines can measure their own performance. In turn, cloud computing makes real-time and historical data available to all stakeholders.
2. Improving Workplace Safety
Manufacturing CEOs and labor unions agree that tasteful applications of robotics and machine vision systems can increase product quality, improve accountability and help workers lift themselves into more skilled and better-paying functions. Just as critically, they’re an ally in the fight for better workplace safety. Machine learning makes collaborative robots a feasible and useful addition to a company’s IIoT infrastructure. Research on cobots finds they improve worker productivity and safety by carrying out and assisting high-risk activities, like reaching for and lifting heavy objects or transporting materials across a busy manufacturing floor.
Today, improved robotic vision means robots have better pathfinding and collision avoidance protocols. Their functionality is not hamstrung by having to work within or behind safety cages.
3. Optimizing Vendor and Supplier Networks
Manufacturers come in two varieties — discrete and process. In the former, you might need a constant flow of parts from multiple parties. In the latter, there might be chemicals or mixing agents involved, without which your core manufacturing process cannot proceed.
The supply chain industry is up against a talent and skills gap as we speak. Let’s take this opportunity to point out that there’s a science to getting all the world’s stuff, including components and finished products, from their source to the end-user. The process is more exciting than most of us might imagine.
Machine learning is increasingly the pacemaker of the supply chain world. With manufacturers using cloud-based enterprise planning systems, it’s almost child’s play to let machine learning-enabled optimization into the mix. The benefits include:
- Predicting when inventories of raw materials will run out
- Predicting and automatically accounting for changes in product demand
- Using web impressions and customer information to inform inventory decisions
- Taking proactive action against weather events and other supply chain disruptions
The supply chain is an area where the right data specialist with the right machine learning tool could wring out a surprising amount of hidden savings. According to one study, new technology significantly improves the accuracy of a demand forecast, reducing errors by 30 to 50 percent. Machine learning and cloud based IIoT networks ensure product shipments are where they need to be to meet customer demand and promote future growth.
Payroll, finances, and facilitating transactions between partners are additional areas where the IIoT, cloud computing, and machine learning blend together nicely. The cloud lets employees and managers manage billing, payroll and and other financial tasks from anywhere, including while they’re working at multiple facilities or on the road. This translates to saved time for management and labor alike.
Meanwhile, machine learning and cloud finances are powering the next generation of fraud detection and compliance technology by automatically looking for questionable records and transactions and flagging signs of fraud, identity theft, vendor malfeasance and more.
4. Making Maintenance More Proactive
All of the manufacturing infrastructures in the world is useless if maintenance intervals aren’t respected. Factories looking to improve their production capacity could recover up to 20% if they take another look at how they apply maintenance on the floor. Today, taking another look means learning how to apply machine learning. Many familiar names, like IBM and Cisco, are staking their futures on building rich networks of connectivity we call the Internet of Things (IoT).
In manufacturing, the IoT gathers data from sensors on conveyors, robotic arms, scales, vision stations and assembly equipment. The IoT solution, using machine learning, figures out what baseline and optimal operations look like. From there, it can flag emerging issues, such as poor performance, as it develops, instead of waiting for failure.
Using systems like these save companies from wasteful preventive maintenance practices, optimizing lifecycles and parts budgets. It also means critical machines don’t have to fall offline before somebody realizes there’s a problem.
Any mistake is costly to make on an ongoing basis. On the global stage, the average cost of equipment downtime is around $5,600 for every minute the machine is unavailable. With machine learning, and with big data powering industrial IoT implementations, downtime should become far less common.
5. Fueling New Types of Design
Finally, machine learning and IoT-connected fabrication equipment make it possible for designers and manufacturers to engage in something called generative design. This step is where materials specialists, product engineers and other decision-makers choose specifications like size and strength, raw material usage and other primary attributes. Then, they use algorithms to generate candidates that meet basic specs automatically.
Generative design — design aided by algorithms operating within user-defined parameters — can solve three problems at once in manufacturing:
- Applying machine learning to the problem of scrap reduction produced savings of up to 30 percent. This technology eliminates waste by optimizing the use of materials without compromising the durability requirements of the final product.
- You can reinvest the corresponding financial savings back into the company. In 2017, only half of the surveyed companies were using machine learning. It is a genuine value-adding asset and a competitive advantage.
- Generative design may reduce testing costs by eliminating tedious trial-and-error and delivering suitable design candidates for digital and real-world testing.
Combined with other technologies like additive manufacturing and the rapid prototyping it unlocks, machine learning will continue to advance the industry in several significant ways.
These examples show that machine learning and the IIoT can complement human ingenuity in some important ways. Manufacturing and distribution are critical enterprises. It’s not surprising to see it’s still a flashpoint for exciting and liberating technological innovation.