Nov_Dec_AMP_Digital

FEATURE A D V A N C E D M A T E R I A L S & P R O C E S S E S | N O V E M B E R / D E C E M B E R 2 0 2 0 5 6 2 PREPARING FOR INDUSTRY 4.0 I n the late 18th and early 19th centu- ries, the world experienced the tran- sition frommanual work to high-vol- ume production enabled by the spread of mechanical machines powered by steam or water. Productivity soared. This became known as the “first” in- dustrial revolution. We are now at the intersection of two newer and more recent revolutions: the third and fourth industrial revolutions. The third rev- olution is known as the digital revolution and is charac- terized by the use of automation, digital computers, and data storage systems. The fourth industrial revolution—cy- ber physical systems—is distinguished by connection to the internet, so-called “smart” hardware, autonomous sys- tems, and machine learning. This revolution is often called “Industry 4.0.” Many of you are already familiar with the home ver- sion of Industry 4.0, commonly called the internet of things (IoT). IoT is the integration of internet connectivity into ev- eryday items such as doorbells, lightbulbs, smart speakers (such as Alexa), and other smart home technologies. The industrial version of IoT is known as the Industrial Internet of Things, or IIoT, which uses network connected devic- es such as sensors and systems, cloud data storage, and machine learning. We are already seeing some early im- plementation of this revolution in the heat treat industry. Many shops now have the ability to monitor and remotely interact with processes through smart phones or tablets. Equipment suppliers to the thermal processing industry are beginning to sell items like wirelessly connected flow- meters and controllers. While these innovative products are beneficial in their own right, the adoption of a more global approach to the IIoT has been slow in many heat treat shops, in part because of the historically conser- vative nature of the business, in part because of security concerns, and in part because the business is extremely cost-conscious (why buy new when the old one still works, right?). For most facilities, adoption of Industry 4.0 is hap- pening in a piecemeal fashion. Even if your company is slow to adopt network-connected technology, integrating the IIoT in your shop should be part of your future plans if you want to remain competitive. Fuller GUEST DITORIAL How will machine learning affect your business? Ma- chine learning is an advanced algorithm that uses statistics to find patterns in massive data sets. It is a multi-faceted concept that can be applied to change the way we control processes. The standard method of control for the past several decades uses the PID loop to tune controllers and provide rudimentary “reactive” control such as avoiding temperature overshoot. Machine learning offers the pos- sibility of true closed-loop “predictive” process control. By leveraging a sufficiently large and organized dataset, algorithms can be developed with the power to predict end-product results using a set of key performance indi- cators (KPI). Each KPI is typically dependent on multiple process variables. This could lead to furnace systems that adapt to process anomalies by making real-time adjust- ments in cycle time, temperature, and atmosphere com- position. Such applications are rudimentary today, but you can expect to see them become more and more so- phisticated and prevalent offering predictive value in the years ahead. How can you prepare your shop to take advantage of machine learning? There are at least three things that you can do today to prepare. First, you should select someone in your organization to develop and manage a facility plan for how you will implement the IIoT. Second, begin buying hardware that is network-capable. Third, make sure that your data is usable. Many heat treat shops already have massive process databases. However, this data is often formatted in a way that will make it difficult or impossible for machine learning algorithms to find meaningful asso- ciations. You should begin thinking about how you collect, format, and store data with future machine-learning appli- cations inmind. If you structure your data properly, youwill increase the likelihood that your data will be usable, help- ing your shop stay competitive in the future. Even though the integration of large-scale machine learning may be a few years away, the time is now to begin collecting “future historical” data. A good step would be speaking with a data scientist about improving how you collect, format, and store data. By collecting well-organized data today, you will be on your way to having the information you need to improve your process tomorrow. Industry 4.0 is here, and it will impact your business sooner than you think. Jeff Fuller Manager Process Quality, Amsted Rail Company Inc.