In a description of the National Science Foundation sponsored center for Smart process Manufacturing (http://www.rockwellautomation.com/resources/downloads/rockwellautomation/pdf/about-us/company-overview/TIMEMagazineSPMcoverstory.pdf) the authors suggest that market disruptions such as a “$3000 automobile or a $300 personal computer” might be outcomes. Plant integration, plant optimization and manufacturing knowledge are listed as the phases to get to this reality. What are the barriers to such an evolution in manufacturing ? How much integration of people, process and technology needs to happen to transform existing manufacturing ? Will leadership for this transformation come from small, agile companies who, when successful, will be integrated into larger ones or can the large companies lead such a transformation ? Finally, how global will this phenomenon need to be to transform supply chains ?
Smart manufacturing refers to using advanced data analytics to complement physical science for improving system performance and decision making. With the widespread deployment of sensors and Internet of Things, there is an increasing need of handling big manufacturing data characterised by high volume, high velocity, and high variety. Deep learning provides advanced analytics tools for processing and analysing big manufacturing data.
With new technologies (e.g. IoT, big data) embraced in smart manufacturing, smart facilities focus on creating manufacturing intelligence that can have a positive impact across the entire organisation. The manufacturing today is experiencing an unprecedented increase in available sensory data comprised of different formats, semantics, and structures. Sensory data was collected from different aspects across the manufacturing enterprise, including product line, manufacturing equipment, manufacturing process, labour activity, and environmental conditions. Data modelling and analysis are the essential part of smart manufacturing to handling increased high volume data, as well as supporting real-time data processing
From sensory data to manufacturing intelligence, deep learning has attracted much attention as a breakthrough of computational intelligence. By mining knowledge from aggregated data, deep learning techniques play a key role in automatically learning from data, identifying patterns, and making decisions. Different levels of data analytics can be produced including descriptive analytics, diagnostic analytics, predictive analytics, and prescriptive analytics. Descriptive analytics aims to summarise what happens by capturing the product’s conditions, environment and operational parameters. When the product performance is reduced or the equipment failure happens, diagnostic analytics examine the root cause and report the reason it happens. Predictive analytics utilises statistical models to make predictions about the possibility of future production or equipment degradation with available historical data. Prescriptive analytics goes beyond by recommending one or more courses of action. Measures can be identified to improve production outcomes or correct the problems, showing the likely outcome of each decision.
With the advanced analytics provided by deep learning, manufacturing is transformed into highly optimised smart facilities. The benefits include reducing operating costs, keeping up with changing consumer demand, improving productivity and reducing downtime, gaining better visibility and extracting more value from the operations for globally competitiveness.
Applications to smart manufacturing
Computational intelligence is an essential part of smart manufacturing to enable accurate insights for better decision making. Machine learning has been widely investigated in different stages of manufacturing lifecycle covering concept, design, evaluation, production, operation, and sustainment. The applications of data mining in manufacturing engineering are reviewed in, covering different categories of production processes, operations, fault detection, maintenance, decision support, and product quality improvement. The evolution and future of manufacturing are reviewed in, emphasising the importance of data modelling and analysis in manufacturing intelligence. The application schemes of machine learning in manufacturing are identified as summarised in. Smart manufacturing also requires prognostics and health management (PHM) capabilities to meet the current and future needs for efficient and reconfigurable production.
- How does smart manufacturing help companies?
- How will smart manufacturing evolve in the future?
By Andrew Gunder, DCMME Graduate Assistant
Edge computing is an emerging trend in the world of technology. Yet, its concepts aren’t completely new. Edge computing is defined as the practice of processing data near the “edge” of your network, where data is being generated, instead of at a centralized data-processing warehouse. It gathers the data at its closest point to make that data actionable in the least amount of time. For example, smart thermostats use edge computing to determine when to adjust temperature through out the day.
Edge computing is beneficial because it moves the computer workload closer to the consumer this reducing latency, bandwidth and overhead for the centralized data center. Content delivery networks (CDNs) are a prime example that showcase the benefits, such as reduced latency and higher uptime, in storing information closer to the end user. It also increases security and reduces risk of breach since the data remains at its point of creation, rather than consolidating in a centralized location such as a server.
Edge computing has emerged as result of the increased prevalence of the Internet of THings (IoT) and IoT devices. The network “edge” relies on use case. Cell towers, smartphones, and automated vehicles all function as micro datacenters for the network. As we continue to see IoT and its related devices expand into mainstream use the limits of edge computing will transcend most physical boundaries on a global scale.
What are some examples of devices that use Edge computing already?
How is the IoT necessary for the success of Edge computing?
How can business and different industries effectively use Edge computing?
Multinational automotive manufacturer, Hyundai, is currently testing an artificial intelligence-based technology that aims to aid hearing-impaired drivers.
Sounds are converted using software and hardware into visual and tactile cues that drivers can read. Sounds patterns are analyzed using AI and sent to two systems: the Audio-Visual Conversion (AVC) and the Audio-Tactile Conversion (ATC). The first system, AVC, converts sounds into pictograms on the head-up display of the car, while ATC transforms sounds into physical cues, such as sending vibrations through the steering wheel.
The first hearing-impaired taxi driver test has been ongoing as of December 2018 in Seoul. Hyundai is in the process of adapting technology to help drivers overcome an impairment, expanding the use of systems from complementing driver experiences to enhancing and promoting driver safety.
How can automotive manufacturing companies benefit from AI?
How is technology solving problems in the automotive industry?
How does AI improve driver safety?
The real breakthrough in the supply chain domain is the arrival of 3D printing as a serious competitor in finished product markets. The technology is slowly gaining acceptance in applications that are “taking it from the prototype to the production-grade stage for smaller components.” The potential changes are many.
Here are a few possibilities.
3PL to Manufacturer 3PL
A new type of 3PL could emerge that offers manufacturing services through 3D printing. Operators like UPS are well positioned to take on this role because a number of intellectual property issues must be resolved before AM methods become ubiquitous. As a trusted third-party provider, UPS has the market stature and scale to function as a new type of hub where products are made, assembled, and distributed.
A New Breed of Agile Supply Chains
With 3D printers operating as standalone installations in strategic locations, companies could manufacture in short runs at multiple sites across the globe. The networks would flex with shifts in demand by reconfiguring the manufacturing nodes or by adjusting machine outputs. Production units shift rapidly from one product variant to another without the need for retooling or lengthy line delays. The AM model also offers tremendous opportunities to cut inventory costs, because there would be less need for inventory. The management of raw materials inventory also would be streamlined as production processes generate less waste.
Armed with 3D printing, machine repair services “don’t have to have every single component; you can print components when needed,” says Ulrich. Positioning parts inventories would become much less of a challenge for teams in the field.
A flexible, highly adaptive network of 3D printing installations could take just-in-time and postponement operations to new levels of efficiency. AM methods could be used to produce precise quantities of customized components very late in the final production cycle when more accurate demand information is available.
New Risk Management Dimensions
Opportunities for improving risk management represent another potential benefit of AM-based manufacturing. Low market-entry barriers and the ability to retool quickly reduce business risk. The technology also provides companies with a rapid-response mechanism when an unforeseen incident disrupts the supply chain.
Since additive fabrication is less wasteful than traditional production processes, it reduces carbon footprints. Similar benefits accrue from innovations such as Oxman’s revolutionary design processes that increase functional efficiency, while reducing material content.
- How are supply chains changing due to 3D printing?
- How does 3D printing help in making the manufacturing industry more green?
- How is 3D printing helping in risk management?
The 2019 Consumer Electronics Show in Las Vegas this year revealed a lot of tech in the car industry. Mostly in regards to self driving cars and improving the user experience by adding new tech in vehicles.
For instance, KIA designed a prototype that adjust the car cabin based on your mood, and will adapt your preferred display panel based on facial recognition.
BMW showed off its self driving motorcycle and backwards tech for 2019 BMW cars. The backwards tech allows you to plan you backwards driving route where margin of error is small. For example, backing out of a tight parking garage with little room for error.
Toyota introduced its Guardian 4.0 Self Driving tech which will help in saving lives on the road. The tech allowed the car to take over the vehicle when there’s an emergency. For instance, if a car pulls out in front of you randomly, the car’s AI will take over the car and react faster than you can to avoid an accident.
Will Toyota Guardian 4.0 tech become a new safety standard for the car industry?
Will the prototypes presented by KIA and other car manufactures become reality in the new feature?
Will there be complete autonomous vehicles on the road or a mix of self driving AI and human operating vehicles?
By Andrew Gunder, DCMME Graduate Assistant
I know what you’re thinking. Dueling AI sounds a lot like a game of Rock ‘Em Sock ‘Em robots, or maybe you’ve conjured up images of the Siri and Alexa arguing with one another. Dueling AI refers to the capability that one AI engine can be used to train another AI engine.
Like humans, AI engines don’t start out smart right out of the gate. They must be educated over time. Take a chess playing robot for example, it may not win its first match or even its second or third, but after studying previous outcomes it begins to recognize patterns and tailor its strategy accordingly. Chess essentially has an infinite number of outcomes after several moves, so it is next to impossible for the AI to learn each move, it must learn instead.
Dueling AI isn’t just limited to robotics. Imagine the ramification on other industries or other possibilities. Suppose we are trying to determine between an authentic image and a digitally altered image. One AI attempts to create a realistic image, and another AI attempts to decide whether the image is real or artificial.
If AI is capable of learning and training another AI, you’re probably wondering what’s preventing it from achieving world domination. Fortunately for us, AI only becomes as smart as the data that it is fed. It possible to “poison” AI by feeding bad data, thus showcasing that as smart as AI can be it is not without limitations.
Can you think of some other tasks AI can train another AI on?
What industries can be impacted the most by dueling AI?
Do you foresee ways that AI might be able to overcome its limitations?
With the increase in accessibility to production and quality data from the use of automation, the Internet of Things, and handheld devices manufacturers are finally able to gather and analyze data to improve their processes at a level hereto unseen before. However, with this seemingly limitless access data comes a new problem: having too much data. More and more companies are falling into the trap of collecting data for the sake of collecting data just because they can and this can actually be harmful to a business. As Douglas Fair states in his article “Drowning in Quality Data: How to Rise Above”, “the insight gleaned from data that is what actually benefits the business”. This means that along with optimizing their processes and machines on the manufacturing floor, manufacturers now also have to think about optimizing how they collect their data so that they are getting the most benefit from it.
When optimization the data collection process, it is important to ask these five simple questions when assessing whether or not they need to be collecting certain pieces of data.
- Why do we need to gather this data? What is the improvement we are trying to make with this data we are collecting?
- How will we use the data after collection? What are we going to do with it after we have collated it?
- Who will evaluate the data? Will it be automated or will we be dedicating personnel to it? Do we have the labor available right now to handle it?
- What is a reasonable amount of data to collect? Can we defend why we need as much as we do or could we do the same thing with less?
- How frequently do we need to collect the data? How often are we analyzing and using the data to make decisions? Do these coincide with each other well?
At the end of the day, the only sure fire way to make sure you don’t fall into “data gluttony” is to check yourself and ensure that you are collecting data for specific purposes, using all the data you collect, and acting on the insights gained from the data to improve your bottom-line.
- With data becoming so centric to operations now-a-days, are we going to start seeing roles dedicated to data analysis on site at plants? How will this affect the way plants are run?
- What are the costs associated with “data gluttony”? Is it really as big a problem as Fair makes it out to be?
- How long does the process of optimizing data collection take? How often should companies review their data collection process to ensure they aren’t collecting useless data?