Step by Step in Maturing Towards the 4th Industrial Revolution
By Dr Tan Chee Pin, Dr S. Veera Ragavan, Dr Chua Wen-Shyan
The 4th Industrial Revolution (also commonly known as Industry 4.0) has taken the world by storm, bringing with it a wealth of opportunities for those agile enough to seize them, but also threats and disruptions to those who fail to recognise them and adapt. For the food & beverage operator, it could mean efficient delivery of their products to capture a market they were never able to reach before. For us, the people on the street, it has changed the way we lived, changing the way we shop (through cashless payments and home deliveries) as well as the way we take public transport. For the manufacturer, it could mean the automation of repetitive tasks, information at the fingertips for quick decisions, and a fully-connected process from the upstream to the downstream, ultimately leading to improved quality of products, shorter production times, and increased productivity.
However, the journey of embracing the opportunities of the 4th Industrial Revolution (4IR) is not always straightforward. With the many options available, one might not know where to start, and there is the risk of making wrong costly decisions by investing in something that might not be suitable for one’s needs.
Based on our experience in developing technologies of 4IR, and working with industries in implementing them, we suggest a step-by-step methodology (see Figure 1), that can be taken in bite-sized and manageable steps, gain some quick wins, until such a time when the industry feels confident to make the investment in the next step. The methodology can be applied to a particular process or machinery (which can be chosen by the business owner, depending on the area of priority). In fact, each step shown in Figure 1, though manageable, could generate immense benefits to the industry, as will be expounded on in the following sections.
Figure 1: Our proposed methodology for adopting technologies of the 4th Industrial Revolution
Data Acquisition
In 2017, The Economist published a story titled “The world’s most valuable resource is no longer oil, but data”, and this holds true even in 4IR. Even in its simplest form, data enables a process manager or business owner to get a quantitative picture of what is going on in the process/machinery, at his/her fingertips through an internet-connected device such as a smartphone or tablet (see Figure 2). For instance, one could know the number of units produced, the duration of operation, duration of downtime, just from this data.
Figure 2: Instant data transmission to process manager/business owner
In moving towards 4IR, it is of great importance to generate a consistent stream of data for various purposes, such as to provide information, for analysis, and to facilitate making informed and timely decisions. It is also possible to make an antiquated machine (that can still run for many years) behave as shown in Figure 2, in a cost-effective manner without having to purchase a new machine. It would firstly involve non-invasive re-engineering and retrofitting the machine with appropriate sensors – this would require a thorough understanding of the machine’s operations and properties. Following that, the sensor data can be pumped to the internet/cloud, and then stored (for future processing) and/or transmitted to a device such as a smartphone or tablet (for instant viewing). Achieving this stage would already be a substantial win, as the data can help a business owner make swift and informed decisions, even though manually.
Data is the world’s most valuable resource of this era
Data Processing
After data has been collected and stored, it can be processed to bring an industry to the next level of 4IR. It can be processed in two ways – firstly to infer the relationship between the variables represented in the data, and secondly to diagnose the condition of the machine in real-time. In the former, the relationship between the variables is commonly known as a model, or sometimes known as the Digital Twin.
These models can be used to simulate the performance of the system for a given set of conditions; users can try various combinations of conditions, to predict the behaviour of the system, quickly, and at no cost nor risk, and make decisions on the settings to be applied. The models can also be used for training purposes, where they can be used to help a user get more familiar with operating the system or to gain more insights about it; this can quicken the training process. The second way of processing data, which is to diagnose the condition in real-time, uses real-time data in conjunction with the model (in parallel) to make an inference of the condition of the system, this is known as condition monitoring. The concept is to inject the process/plant/machine input into the model, and compare the outputs of the model and the process/plant/machine, and the discrepancy is further processed (using artificial intelligence or machine learning algorithms) to make a diagnosis (see Figure 3 for an illustration of the concept).
Figure 3: Schematic figure of model-based condition monitoring
This diagnosis can automatically generate an alarm if an abnormal condition is detected, or predict the future performance of the system and estimate when it will break down, which is known as predictive maintenance – that can help a business owner make plans on maintenance or replacing the machine. Finally, online and real-time data can be used in a feedback loop to make automated decisions.
Decision and Action
After data (and/or diagnosis) of the process is available, then decisions/actions can be taken, either manually or automatically. For example, if a machine is diagnosed to be behaving abnormally, the process manager can decide to either proceed with production and schedule repairs later on, or to halt it immediately (as the abnormality could reduce product quality or damage the process); this is a manual decision. An example of an automated decision is the machine-based adjusting of an input (such as a valve opening or motor voltage) in a feedback control loop, where the decision could be derived based on various methods such as control theory or artificial intelligence. Figure 4 illustrates this concept, where the condition monitoring block in Figure 3 could form part of the Controller/Decision block in Figure 4.
Figure 4: An illustration of an automated decision
The next level of actions, that are more complicated, could be new ways of actuation or handling products. For example, this could be the design of a new robot or mechanism to handle textiles/apparel, or to handle odd-shaped or delicate products such as tomatoes and eggs. In this situation, there are many more possible scenarios or configurations to be considered, and the development of these kinds of mechanisms could be much more costly.
Conclusion
In summary, 4IR offers many opportunities for an industry to move to the next level of productivity, and we have described a step-by-step methodology to achieve it. Each step can be executed separately, offers immense benefits on its own, and builds on what has been achieved in the previous step. In this way, the business owner can incrementally improve on the development of his/her processes at a comfortable and manageable pace, depending on the financial resources available and constraints imposed (without having to make an upfront huge financial commitment to implement all steps at one go), and reap the benefits along the way whilst gaining the confidence to move on to the next step.
Dr Tan Chee Pin (Edwin) is an Associate Professor at the School of Engineering, and heads the Robotic & Mechatronics Engineering program at Monash University Malaysia. The program heavily involves 4IR technologies, both in teaching and research. He is a member of the Conference Editorial Board of the IEEE Control Systems Society.
Dr S. Veera Ragavan worked for several multinational companies in various capacities from a Design Engineer to Business Unit Head. He has more than 17 years of industrial experience in design and development of Factory Automation Systems and has executed several projects from concept to commissioning. He is currently a Senior Lecturer in the Robotics & Mechatronics Engineering program at Monash University Malaysia, training many students and helping many industries in moving towards the 4th Industrial Revolution.
Dr Chua Wen-Shyan is currently the Head of Malaysian Smart Factory 4.0 under the Selangor Human Resource Development Centre (SHRDC). He is also appointed by the Malaysia Productivity Corporation (MPC) in collaboration with the Machinery and Equipment Productivity Nexus (MEPN) as one of the mentors (advisors) for the PRODUCTIVITY1010 initiative to support the industries in their journey towards digital transformation.
