Navigation Tools for Smart Factories…!
Not every industry that is automated with software, is a smart factory. Not every industry that is connected with sensors for data collections, is a smart factory. Not every industry that is performing big-data analytics, is a smart factory. Not every smart factory is really smart as it appears (https://medium.com/@alam.hilaal/how-smart-is-the-smart-factory-323a9ea8131f). A true smart factory is the one which has enabled informatics to achieve its objectives and vision…!
Old Wine…
During the pre-digital age, manufacturing industries adopted many innovative process tools and techniques to be efficient, in other words “smart”.
Quality Function Deployment (QFD)is one such tool that has been used by the the product designers, especially in the automobile industries in Japan, USA and European countries. Despite of its powerful approaches, there were reluctances shown due to its series of complex matrices and a large amount of information for integration. QFD requires upfront investments in the beginning of the project phases and this annoyed many especially CFOs. Moreover, despite of its rationality, the decision making processes are both exciting and painful, requiring a long preparation time.
QFD like many other tools and techniques, needs to maintain a large amount of data and information from various departments in an industry. Integrating them precisely is a very challenging part and this process generally discourages the team in due course of time. I have seen many QFD matrices being maintained in spreadsheets and the houses of quality (HoQ) with CAD drawings. There is no useful software except two or three and they also have severe limitations.
In addition to QFD, there are other design and process management tools such as FMEA, FTA, TRIZ etc. which are widely being used in many industries, having the similar challenges. The crux of the matter is, the management tools need accurate information and data for a perfect navigation.
Subjective to Specific…
The greatness of QFD is its ability to facilitate the transformation of subjective information into specific details. During this transformation, there are various stages of decision makings and judgements. In order to enable these stages effectively, reliable sources of information are essential.
The difficulty in data handling in QFD causes the big projects to be broken into several small chucks. Many a time this approach prevents a creative integration of two (or more) subsystems into one, i.e one subsystem may be sufficient to solve several other tasks; however, by breaking into chucks results in “excesses”. This also could be one of the major reasons for not adopting it widely.
Dynamics of Manufacturing Industries
The manufacturing industries at present are not like in the past. Unlike in the 70s, when QFD was invented, the face of the manufacturing techniques is changing fast in phase with other modern technologies. In addition to the fabrication processes, informatics have been put in more emphasis in the modern day machines. This lead to the realtime data acquisition technologies to emerge with internet era, and thus, making a baby step towards the age of automation.
Then came robotics to automate many processes. For two decades, robotics were operated with localised control systems. The convergence of internet and robotic control systems enabled several remote operations and paved path for remote collaboration among robots.
Be it a traditional fabrication or 3D printing, accelerating MEMS sensors developments caused synchronized operations among the service robots and the robotic-machines, changing the landscape of manufacturing industries once for all in this era.
Industrial IOT
Internet of Things (IoT) of some standalone devices steered towards the development of microcomputers such as Arduino, Raspberry PI etc., so that hobbyists could make use them. Taking cue from there, real serious developments were started in home automation, automobile sensors etc. In parallel, industries which had adopted intelligent control systems, applied IoT in their production processes scripting the new chapter — IIoT.
Today Industrial IoT (IIoT) is all about continuous stream of data from every nook and corner of the whole industrial eco-system. Low cost MEMS sensors, high bandwidth of communication and low cost data storage made way for IIoT. The machine health, machine performances, their output, product details and their runtime data ensure the quality of the products & processes with this new technologies. Currently IIoT is mainly used for predictive maintenance and processing internal operations remotely. Main stream software applications and modern machines are being integrated in order to monitor the entire production processes in an industry.
These steps ensure the reliable data from various stages and phases of an entire process. As of 2017, there are around 28 billion IoT devices has been installed around the world and expected to reach 35 billion by 2018. The IoT technology is flowing towards the convergence of “Things — oriented”, “internet — Oriented” and “Semantic — Oriented” orchestrating the entire eco-system for smooth flow of operations. These data can be well-used in management tools for the subsequent action plans and they can even be automated…!
Artificial Intelligence & Machine Learning
Machine leanring (ML) technology bundles the data available to it and makes decisions based on the previous “experiences”. ML has thus become a part of the artificial intelligence (AI) comprising the “deep leaning” in it. Unlike the traditional programming techniques which spit outputs based on inputs, ML is “creating” new programs from the inputs based on its leaning capabilities. A computer program is said to learn from experience (E) with respect to some class of tasks (T) and performance measure (P), if its performance at the tasks improves with the experiences.
AI is a system that “behaves” like human beings. It uses “Agent” that can gather information about its environment and take action based on those information. Reactive agents are rule based systems that needs no memory setup. The second type of agents — ML agents — take information from memory where experiences are stored. ML feeds necessary information and data for the second type of agents to make decisions. The term “SMART” enters when IoT and AI comes along with “economy”. Any system can be called smart when these three are integrated in order to achieve the mission and vision of a process.
Every stage in QFD can be made taught to the systems and thus, the human intervention can be minimised and automated in a longrun. AI technology can be used to make predictive priority decisions and design selections based on conditional rules and machine learning algorithms. Inclusion of ML & AI in QFD would enormously enhance the effectiveness of the product development processes and manufacturing. Integration of IIOT, ML & AI with QFD would script a new era in manufacturing technology.
Blockchain Technology
Blockchain is the latest wonder kid of information technology in this millennium. It is a distributed database at several nodes instead of maintaining them at a centralised location. The update process uses the cryptographic techniques and hence, it is still not hackable. Implementation of blockchain begins with the simple questions like:
Does the double (/ multi) spending (using the same proofs or copy and paste) affect your business?
Need somebody to authorise which causes delay? (eg. Set of instructions / contracts)
Does the intermediate consume your profit share / cause delays?
Fear of cyber — hacking?
Fear of calamity or destruction?
Internet security is a major concern now-a-days. The same goes to IoT too. The consequences of any compromise in security of the system will result in unimaginable disasters for industries and economy alike. However, the emergence of blockchain technology from the different quarter may help mend this problem. Integrating blockchain with IOT systems is likely to ensure the security of the entire ecosystem though it is not well-discussed or analysed yet.
… in the New Bottle…
Emergence of artificial intelligence, machine learning, MEMS technology, blockchain protocol join together to give life back to many abandoned excellent tools and techniques with fresh outlooks. These tools and techniques will put the industries on structured roadmap for navigating for technological adoptions rather than simply adopting ML, AI, IIOT & blockchain in a specific area or with no or little purposes.
Even though cloud computing technology has not been discussed here, it has been the enabler to connect all other technologies, and functions as the central nerve system of the entire ecosystem. This will pave path for the creation of futuristic revolutions such as circular economy, sharing economy and automation of everything.
