Digital Twin and IoT
In the age of data, everything seems to be digitized in a way. Digitization affects most plants and factories. They cannot escape from the power of data and digitalization instead everyone must use it for their good. Right now, the most promising technology is cloud computing. With cloud computing, everyone can have enough computer power at a relatively cheap price. Cloud computing is efficient, accessible, cost saver, secure and flexible [1]. Also, other technologies can be made accessible with Cloud Computing. There are new technologies that can effectively use with cloud computing. Big data analysis and storage are becoming the superpower of factories. To store terabytes of data, big processing machines are needed. However, cloud computing makes it accessible to everyone. Big data can be used in the Automation and Enterprise Platforms, factories, plants, and manufacturing areas. Blockchain technology enables the usage of technologies with no security issues. 5G ensures reliable, fast data transfer between platforms, enhancing and speeding up the use of new technologies. These technologies found their way into the workforce and factories. AI, Machine Learning and applications, distributed ledger, advanced data analysis, etc. created new opportunities and income channels. With the increased usage of sensors and a huge amount of data processed, engineers tried to harness the power of data. Data returned them with more power and efficiency After using more sensors and data, The Internet of Things (IoT) was born which is the subject of this article.
Radio Frequency Identification (RFID) System is the origin of IoT technology. An IoT system is composed of a reader that reads the information with a wireless system. The reader is a device with one or more antennas that emit radio waves and receive signals from the RFID tag. IoT essentially is the series of connected RFID systems to a terminal and HMI which shows data in real-time. IoT is very useful in dangerous situations, high-temperature processes, future prediction, split-second decisions, and in cases where real-time tracking is important. It is powered by numerous sensors, processing abilities, and software.
The area which will be investigated in this article is IoT in factories. Real-time monitoring and decision-making helped engineers to constantly monitor equipment, production, and maintenance status. Engineers understand that production and maintenance can be more deficient-free and efficient with the IoT and constant monitoring. Monitoring enables the prediction of machine deficiencies before it happens. Defects can be seen and acted on it in real-time. Along with real-time decision-making, predictive maintenance is a benefit of IoT. According to the study by Oden Technologies [3], predictive maintenance reduced the unplanned downtime by %3.01. By constantly monitoring equipment status, one can predict the exact time when equipment needs maintenance, which can save money and time that otherwise will pass with no production.
IoT with a lot of sensors and data can create a virtual version of a physical object that resembles the actual object. This can be used for simulation and prediction. Sensors, providing real-time data enable engineers and supervisors to monitor crucial steps on the shop floor. Sensor usage and real-time monitoring, a digital copy of the plant can be created and can be monitored on an HMI. A “Digital Twin” of a plant. This “Digital Twin” will have the same parameters as the physical object, sensors can monitor the process in real-time. A virtual twin of the physical object. “Digital Twin”.
A digital twin is the virtual model of a physical object where you can monitor and change the status of the processes. As can be seen in Figure 2, the Digital Twin can be created by merging data relevant to the process and critical parameters. Data are used to create a virtual model of a physical object. To create data from the source, various sensors are used. Pressure, temperature, etc. can be monitored. However, sensor choice only depends on what is important in the process. By adding control objects to important places (it can be added to sensors), monitored parameters can be controlled manually or automatically. Although digital twins are used for monitoring and controlling, there is a wide range of areas that it’s used. Better R&D and simulation of hypothetical processes can be achieved without actually using the equipment. With monitoring, users can make split-second decisions about intermediate and end-product. Simulation of real-world cases is a plus for the digital twin. With a digital twin of the same system, the new system can be up and running in a short time. Because of the known processes related to the system and digitized version of the process, slow documentation and human error can be prevented.
IoT is not a cost-saver for every process. If there is no need to monitor and watch the temperature of a tank it would be inefficient and costly to have a temperature sensor. Also, if the object is not complex enough, it may be unnecessarily costly for implementation.
To learn the IoT and digital twin of a physical object, one must learn the data flow of the sensors. Data flow is; constantly monitoring the status of a physical object, data relay to a processing system, and applying to the digital copy [4]. The most beneficial objects to having a digital twin are [4];
· Physically large objects
· Mechanically complex objects
· Power equipment
· Manufacturing projects
Physically large objects can be controlled and monitored efficiently if it has high-level engineering inside. Mechanically complex objects require various sensors and monitoring which IoT can help. Complex objects especially become powerful with IoT because of big data producing and processing. Power equipment outputs and inputs can be monitored by flow sensors, outputs can be controlled easily. By measuring inputs and outputs optimization calculations can be made easily, and efficiency monitoring and control can be made. Manufacturing projects require efficiency and increased productivity. Digital twins, with the help of people, can streamline and optimize manufacturing processes.
The digital twin creation process for engineers is not an easy task. But with the sequenced, systematic approach, factories can adapt to new changes easily. A digital twin can be created in sequence;
1. Designing physical objects on the virtual medium at CAD or 3D modeling.
2. Data processing for design decisions.
3. Simulation of the physical environment on the virtual medium.
4. Calibration of a physical object for calibration of the digital twin.
5. Connection between physical systems and cyber systems in real-time.
6. Data collection for continuous system integration.
The most important step is the 4th step in this sequence. Calibration of the physical object which will be adapted to the virtual twin is a very complex process for engineers. The whole manufacturing process can be changed to implement sensors. Calibration of the process does not end with changing the manufacturing process, sensor places can be changed and optimized. Sensors can be put in various places to acquire reliable data. These must be thought of to get reliable data.
Data is nothing without a processor. It can be a human or computer that is processed but it must be refined and processed to have meaningful results. Processors refine and process the data to make data readable to humans. Machine learning and AI can be powerful tools for data analysis. Already AI’s such as PowerBI Insight gives data analysis to analysts which helps people to analyze. AI and Machine Learning will be the key to refining, analyzing, and giving insights from data.
Although some AI gives insights to processes (such as PowerBI Insights) for optimization and decision-making, a human must be the decision maker in double-edged situations. AI can be cold-blooded when it comes to human life. AI can think quantitative, not qualitative. Because of this, when a complex situation with various parameters occurs, a human must be the decision maker.
But there will be another side to this subject. AI and Machines can and will explain their reasoning always. There will be inputs and outputs always in the case of the machine. However, sometimes inputs are not clear for Humans, even to themselves. Humans cannot defend themselves when they do not know their motives and emotions while taking decisions. This will be the key difference between AI and Humans. And the future will decide which way will be chosen.
AI helps humans to decide. Some non-important processes can be decided by the AI, of course. However, human interaction is needed for complex and double-edged decisions, right now. AI will get better in data processing and suggestions. This can be helpful for big data processing and digital twins. The human workforce and qualified personnel are important in every manufacturing plant and process. Consulting is very important for steering plants and factories for smart applications. I think consultants of the technology firms are and will play an important role in steering smart applications. However, consultants of the technology firms have limited knowledge to implement improvements in manufacturing such as lean manufacturing. To measure and calculate KPIs related to the efficiency of machines and personnel, IoT and MES must be used by qualified consultants. The factory of the future will require qualified integrated production management systems that must be accompanied by a consultant who is just as qualified to achieve its full potential.
Constant data flow with a processor creates insights along with controlling and monitoring. Long-term planning and predictive maintenance can be more realistic and predictive with the usage of Digital Twin. With the processing capabilities of cloud computing, computers can give suggestions and create relations humans would never think of.
IoT sensor prices are continuing to drop every year. This is an important factor when a company decides to build a digital twin. Also, sensor security is a very important subject that can be exposed to cyberattacks. Every sensor in the process is a potential candidate for a cyberattack and needs to be secured. Data needs to be secure and stored. Past data must be accessible. In the case of pharmaceutical companies, it must be guidelines for the IoT and compliant. Computing capability is where cloud computing comes in. Cloud computing can be a solution for both storage and processing capability. Cloud computing provides necessary resources for data analysis. Data sets are needed for reliable data analysis. A digital twin can help with the categorizing, visualizing, and contextualizing of the data. Related activities can be filtered and can provide meaningful results to users.
According to Gartner, there will be 20.4 billion IoT devices by 2020, creating more than 500 zettabytes per year in data [6]. The world is changing with data usage and sensors. Plants and factories which cannot keep up with the changing world are destined to be outdated. Plants are already digitized by the usage of ERP. ERP plans and manages the resources of the factory while MES controls and assures quality in the manufacturing processes. Digitization already started. According to the American Society for Quality (ASQ) [7], based on surveys on 2014, by using the IoT, %82 companies increased their efficiency, %49 companies had fewer manufacturing errors, and %45 companies increased their customer happiness.
When done correctly, a digital twin might help with theoretical situations, simulations, and meaningful predictions. By combining digital twins with AI, one can create a future asset [6]. AI, which will be perfect at interpretation, will help users with the explanation and meaningful results. With all these technologies, one can create a plant that belongs to the future.
References
[1] “Why is cloud computing important? — Open Cirrus.” https://opencirrus.org/cloud-computing-important/ (accessed May 10, 2022).
[2] “IoT Çözümleri — Destech Internet Hizmetleri.” https://destech.com.tr/hizmetler/iot-cozumleri-internet-of-things/ (accessed May 10, 2022).
[3] “What is a digital twin? | IBM.” https://www.ibm.com/topics/what-is-a-digital-twin (accessed Jun. 02, 2022).
[4] “What are digital twins and how are they used in industrial manufacturing?” https://www.motioncontroltips.com/what-are-digital-twins-how-are-they-used-in-industrial-manufacturing/ (accessed Jun. 02, 2022).
[5] “Reliabilityweb Digital Twin: Transforming Asset Operations.” https://reliabilityweb.com/articles/entry/digital-twin-transforming-asset-operations (accessed Jun. 08, 2022).
[6] F. Shrouf, J. Ordieres, and G. Miragliotta, “Smart factories in Industry 4.0: A review of the concept and energy management approached in production based on the Internet of Things paradigm,” IEEE International Conference on Industrial Engineering and Engineering Management, vol. 2015-January, pp. 697–701, 2014, DOI: 10.1109/IEEM.2014.7058728.
