Lean and Six Sigma in the Age of Smart Manufacturing

Lean dates back to the 1960s when Toyota introduced Lean in its Japanese production plants. Because computer systems were primitive or nonexistent on the factory floor, Lean philosophy relied on simple visual controls. Lean also empowered all employees to make decisions that could shut down an entire assembly line, a far cry from US practices of the time.

Six Sigma dates back to the 1980s with Motorola followed by GE. It provided a commonsense framework for solving complex problems using the scientific method. Six Sigma data analysis drives the effort to reduce defects, improve quality, and optimize operational efficiencies.

Most organizations have combined the philosophy of Lean with the problem-solving framework of Six Sigma. Many organizations have rebranded their Lean Six Sigma programs as Continuous Improvement and more recently as Industry 4.0 programs.

With Smart Technologies, Lean and Six Sigma have a new lease on life, becoming more efficient and more effective. One of the best ways to envision the change is to picture the traditional process of an industrial engineer or continuous improvement team member watching a manufacturing process and noting cycle times with stopwatch and clipboard.

Imagine the ineffectiveness of trying to accurately capture the variations in a physical operation across various machines, across three shifts, and across each day of the week and each season of the year. Regardless of the skill and dedication of the analyst, they can only observe and document a small percentage of the entire population of operations. Now imagine smart cameras and IoT sensors watching all transactions on an Edge computer (a computer located near the action) and transmitting data to the Cloud for analysis.

Exhibit 1.7 shows a worker making notes on clipboard, the traditional method of data collection on the factory floor.¹⁵

Smart Manufacturing offers quality and process improvement professionals robust digital tools to examine and evaluate operations without the labor-intensive and ineffective practices of the past. AI and computer vision provide the means to automate visual inspection with greater accuracy and consistency than using manual methods. The new technology also provides data sets for all transactions, not the small sample sizes used in the past.

Exhibit 1.8 demonstrates just how small a sample size is required to meet the Military Standards that have been in place since the 1950s. In this example a 500-part sample is less than 2% of the total population.¹⁶ With smart cameras and IoT sensors watching the action on a 24/7 basis, the new sample size is the entire population of 35,163 parts. The combination of Edge computing and the Cloud provides an easy means to run statistical analysis leading to improved quality.

Improving Cybersecurity Using Smart Technology

Cybersecurity threats are coming at organizations from a variety of sources, including those sponsored by foreign governments hostile to Western democracies, and from criminal sources, both foreign and domestic. What they have in common is a very successful track record of overcoming firewall protections to steal and hold hostage critical company information and cripple operations.

It is a big mistake for manufacturers, especially smaller ones, to believe that they are not a cyberattack target. They are, for the simple reason that they are easy to breach. Here are 10 key cybersecurity takeaways to consider:

1. Globalization, specialization, and IoT trends have increased cyber risk.
2. Supply chains are much deeper and broader than you realize.
3. The supply chain is an attractive target for several reasons.
4. No supplier is too insignificant to be immune from risk.
5. Security controls are only as strong as the weakest link.
6. Threat actors have a wide range of motives and methods.
7. Cyber risk can be mitigated by making business tradeoffs.
8. Impact of a breach can be mitigated with proper controls.
9. Cyber risks need to be considered in sourcing decisions.
10. The costs of a breach can be far-reaching and catastrophic.

Smart Logistics

The modern science and practice of logistics had its origins in World War II. American logistics practices were a primary factor in the Allied victories over Germany and Japan. Logistics is the process of managing the end-to-end planning, acquisition, transportation, and storage of materials through supply chains.

Logistics 4.0 revolutionizes the practices that help win wars and power modern manufacturing and distribution. The digitization of logistics operations includes driverless trucks, delivery drones, automated warehouses, smart ports, smart containers using radio-frequency identification (RFID), blockchains, and AI-powered routing of parts. Smart Logistics come at a critical time to help mitigate supply chain shocks from pandemics, tsunamis, trade wars, shooting wars, and the instability inherent in less developed economies. Finally, Smart Logistics may be the only option to solve the chronic shortage of local and long-haul drivers.

Big Data for Small, Midsize, and Large Enterprises

Smart Manufacturing's ultimate goal is to digitize all physical operations, creating a constant stream of data in real time, typically captured on Edge computers and communicated to the Cloud. Big Data is not a goal of Industry 4.0 and Smart Manufacturing. Manufacturing organizations have generated large volumes of structured and unstructured data for several years. The problem is that much of the data ends up in silos, not extracted or normalized for analysis.

Smart Manufacturing is transforming traditional manufacturing by replacing isolated and siloed data with the ability to collect both structured and unstructured data from diverse sources. Therefore, the goal of Smart Manufacturing is to mine, merge, and transform data to provide a digital twin of operations in real time. Without Big Data analytics much of Industry 4.0's technology is wasted, just as large amounts of data were wasted with Industry 3.0 technology.

Big Data makes possible predictive modeling that exploits patterns found in historical and transactional data to identify risks and opportunities. While some of the most advanced Big Data technology solutions may beyond the reach of smaller organizations, there are many affordable and easy-to-use Big Data tools that smaller organizations can utilize. Today's global supply chains are dynamic, with multiple levels of dependencies. Without Big Data, it is not practical for an organization of any size to adequately identify risks and opportunities.

Industrial IoT Sensors

Smart sensors are devices that generate data transmitted to Edge computers and to the Cloud to monitor various processes. They are typically easy to install and require little configuration. The types of sensors are quite varied:

• Temperature sensors
• Pressure sensors
• Motion sensors
• Level sensors
• Image sensors
• Proximity sensors
• Water quality sensors
• Chemical sensors
• Gas sensors
• Smoke sensors
• Infrared (IR) sensors
• Acceleration sensors
• Gyroscopic sensors
• Humidity sensors
• Optical sensors¹⁷

With Smart sensors, data is automatically collected and analyzed to optimize operations, improve safety, and reduce production bottlenecks and defects. Sensors communicate data to Edge computers and/or the Cloud via IoT connectivity systems on the factory floor. IoT technology leverages wired and wireless connectivity, enabling the flow of data for analysis. It is now possible to monitor operations remotely and make rapid changes when warranted by conditions. The use of Smart sensors helps improve manufacturing processes and product quality while reducing waste and safety violations on the factory floor.

Exhibit 1.9 shows the flow of data from several types of IIoT sensors to Edge computers for analysis and to data monitoring applications and dashboards.¹⁸

Artificial Intelligence Machine Learning and Computer Vision

Today's computer vision has the goal of helping computers see. It uses artificial intelligence and machine learning to digitize imagery for analysis. Tasks that come easily for humans are a challenge for computer vision. A human easily understands that a car in the distance moving toward them appears larger as it gets closer. Computers need to be taught that the change in size does not indicate several different cars. I recall the early days of my supporting a computer vision startup. The engineers were excited that they had taught their program to detect a bare arm reaching for a controller. It worked fine until someone wore a long-sleeve shirt. The program ignored the arm because it had not been taught to consider an arm with clothing.

The manufacturing use cases for computer vision are varied and continue to grow. As the affordability and ease of installation continue to improve, it is reasonable to predict the demise of “dumb” cameras, even for consumer uses. Some of the more popular computer vision (CV) applications include:

• Human/Machine Interaction. CV is able to capture how operators interact with their equipment. Unlike manufacturing shop floor control software programs, CV captures the actual efficiency of operators (the number of units produced against a standard) and utilization of equipment (the hours in operation against the total available shift hours). All this is accomplished without entering data or scanning barcodes. CV can also identify the actual percentage of time an operator is at their workstation versus away from it. This is also accomplished without entering data or scanning barcodes, so there are no data entry errors.
• Anomaly Detection. CV is able to flag such anomalies as forklift and truck drivers speeding, people standing in restricted areas, an excessive number of workers congregating in a work center, or conveyors running without materials or parts after an established maximum time.
• Defect reduction. CV is able to automate visual inspection to complement or replace human inspectors of parts, assemblies, and packaging. This overcomes inconsistencies from inspector to inspector and from shift to shift. Multiple smart cameras capturing 100 images per second can inspect even the most complex assemblies in a few seconds.
• Barcode and Label Scanning. This is one of the simple use cases for CV eliminating human error and inconsistencies.
• Safety and Security Violations. CV is able to help organizations maintain social distancing and mask requirements during the COVID-19 global pandemic. In some cases, CV has helped to prevent serious injuries when operators clearly violated safety requirements by standing behind trucks in busy terminals or not wearing their safety harnesses and hard hats in restricted areas.

Networking for Mobile-Edge Computing

Smart Manufacturing is powered by mobile computing. Without mobile technology, Smart Manufacturing would not be practical. Mobile devices are the platforms by which manufacturing workers and managers can connect easily to the Cloud. The IIoT generates massive amounts of data with connected devices. By combining mobile's ability to provide networks with the data generated by the IIoT, manufacturers have powerful new sources of information to improve operations and eliminate paper-based practices.

Mobile communications has been with us for decades. The first mobile communication was designated as 0G and generally thought to start with the car phone, introduced in 1946 by the Bell System. Beginning in the 1980s, the first generation of wireless analog cellular phones was introduced. These are part of the 1G generation. Starting in 1991 in Finland, the first commercially available digital cellular phones were introduced, creating the 2G generation. Beginning in 2009, 4G was commercially launched in Sweden and Norway. In the United States the launch was in 2010. Then in 2019, 5G technology was launched almost simultaneously in South Korea and the United States, and the rollout is expanding now around the world. Finally, 6G is expected to launch commercially sometime after 2030.

5G has provided manufacturing with the high bandwidth, low latency, and high reliability that are critical to many mobile computing applications. In the past these applications required fixed-line connections. 5G technology will be key to increasing flexibility, shortening lead times, and lowering costs on the factory floor. While 6G is still years away, some early estimates predict 10 to 100 time increases in speed. This will continue to drive down costs while increasing the capabilities of all types of mobile computing devices.

Edge computing refers to the location of a computer relative to sensors feeding it data. Exhibit 1.10 is a graphic showing the process of gathering data from smart sensors and transmitting it to Edge computers for real-time alerts and actions and then to the Cloud for analysis.¹⁹

Additive Manufacturing and 3D Printing

Additive manufacturing, often referred to as 3D printing, plays an important role in Smart Manufacturing. It is especially useful for small production runs and rapid prototyping, which helps shorten the times required for new product introduction (NPI) while lowering development costs. It also allows for greater product customization by eliminating the long setup times required with traditional production processes.

3D printing makes objects according to a 3D digital model. 3D printers add material layer by layer, creating objects ranging from simple to highly complex. Originally limited to plastics, 3D can now use a variety of metals to produce parts quickly and inexpensively. As this technology advances, 3D printing can move beyond prototyping and short runs to support agile manufacturing, allowing consumers to expect multiple design changes over short periods of time.

Exhibit 1.11 shows a 3D printer creating a complex object.²⁰

Robotics

Robotics will play a central role in Smart Manufacturing and will grow in importance as its costs continue to decline. Many workers fear that robots will replace them. The truth is that robots will help to save manufacturing jobs by keeping domestic producers competitive with foreign competitors who enjoy lower-cost labor. In many cases robots are performing dangerous or repetitive tasks so boring that nobody wants to do them. This frees workers to focus on higher-value and more interesting activities.

Some of the benefits of manufacturing robotics include:

• Improved efficiencies throughout all phases of production
• The ability to operate on a 24/7 basis and in a lights-out environment
• Greater agility to react quickly to product and process changes
• Increased ability to bring back domestic manufacturing that was previously offshored
• Help in addressing chronic labor shortages, especially for boring, repetitive, or dangerous tasks
• Improvements in quality and reliability
• ROIs achieved in a few months in many cases

Exhibit 1.12 shows a small robotic arm with a mechanical hand.²¹

Smart Technology to Improve Life on the Factory Floor

Over the past 50 years the number of manufacturing jobs in the United States has dropped from 20 million to 12 million. Much of this has been attributed to automation, according to the Center for Business and Economic Research at Ball State University. Of course, offshoring plays a major role as well.²² Unfortunately, this has given Smart Manufacturing a bad reputation among American factory workers.

What is missed in the criticism is the great number of Smart Manufacturing tools that help make life better for factory workers. Inexpensive smart sensors monitoring production equipment and tools can spot problems and prevent plant shutdowns. Such disruptions reduce revenue and customer satisfaction, potentially leading to loss of wages and layoffs. IIoT sensors and computer vision can also help to flag safety and security violations reducing accidents and injuries.

Adding smart sensors along the entire production line provides a wealth of valuable information that can foster continuous improvement and Lean Six Sigma programs. The new data sources do not necessarily make equipment more powerful, but they do make humans more knowledgeable. This leads to better decision making.

Improved Quality and Safety

The cumulative effect of IIoT sensors, computer vision, robotics, smart machines, data analytics, and so forth will help to reduce errors and defects by eliminating the bulk of inefficient human interactions. It will also help to standardize and automate processes around best practices. These same technologies will help flag unsafe activities on factory floors and in distribution centers.

Improved Margins

Robust new data sources are creating a digital twin of operations so manufacturers can determine optimal areas for automation. In the past, incomplete data did not accurately capture physical operations, leading to investing in new equipment that was not needed or investing in the wrong types of equipment.

Improved Cycle Times

Lean has always striven for continuous improvement to reduce manufacturing cycle times. Smart Manufacturing identifies new opportunities to reduce cycle times. AI makes machines intelligent, robots shorten production times, and computer vision identifies production bottlenecks. The combined effect is to shorten cycle times without sacrificing quality or safety.

High Efficiency with Well-Defined Smart Factory Processes²³

Lean also strives to replace traditional material systems that pushed production through factories based on forecasts with a pull system based on customer orders. This is only possible after the great majority of waste is removed from the entire manufacturing cycle. What results is the ability to make to customer orders, even if the lot size is only one. For 35 years in manufacturing this seemed an impossible dream in the factories I supported. With Smart Manufacturing, this dream is now achievable.

1. What are the major differences between Smart Manufacturing and Industry 4.0?
   1. Industry 4.0 is only used in Europe.
   2. Smart Manufacturing is only used in the United States.
   3. Smart Manufacturing generates more data than Industry 4.0.
   4. They are different names for the same thing.
2. A major change from the First Industrial Revolution to the Second was the addition of
   1. Steam power
   2. Water power
   3. Electric power
   4. Computers
3. A major change from the Second Industrial Revolution to the Third was the addition of
   1. Computers
   2. Electric power
   3. The assembly line
   4. Modern highways
4. The Fourth Industrial Revolution, or Smart Manufacturing,
   1. Creates a digital twin of physical operations
   2. Relies heavily on mobile computing
   3. Deploys many types of IIoT sensors
   4. All of the above
5. Computer vision helps to improve safety by
   1. Creating alerts to safety violations in real time
   2. Requiring managers to review hours of video
   3. Preventing workers from violating rules
   4. Automating tasks
6. The benefits of robotics include
   1. Improved efficiencies throughout all phases of production
   2. Ability to operate on a 24/7 basis and in a lights-out environment
   3. Greater agility in reacting to product and process changes
   4. Increased ability to bring back domestic manufacturing that was previously offshored
   5. All of the above
7. Computer vision
   1. Can only analyze vehicles
   2. Does not typically require artificial intelligence (AI)
   3. Digitizes physical operations to improve efficiencies and safety
   4. Is too expensive for most organizations
8. The role of Big Data analytics in Smart Manufacturing
   1. Is critical because of the many new sources of data created
   2. Is of minor importance
   3. Requires PhDs to analyze the data
   4. Can only be used by large organizations
9. What is Smart Manufacturing's (SM's) role in creating high-paying jobs?
   1. SM is a jobs killer.
   2. SM automates low-skilled, mundane, and dangerous jobs.
   3. SM creates high-paying jobs by helping to keep American manufacturing competitive.
   4. SM has no role.
   5. b and c
10. How does Smart Manufacturing improve quality?
    1. It uses robotics to eliminate errors and standardize operations.
    2. It digitizes all process steps for analysis – there is no need to use small sample sizes.
    3. It uses computer vision to quickly accept or reject parts.
    4. All of the above