Industry 4.0 in India- A global perspective
Sundarram is a business leader with a B.Tech in Mech- anical Engineering from IIT Madras and an MBA from IIM Calcutta, and has held leadership positions in well-known companies. Sundarram served as President of Rane Brake Linings from 1999 to 2004, leading them to receive the Deming award for their successful Total Quality Management (TQM) journey. He has also held CEO roles at ESAB India and Cosmo Films.
In a conversation with Charulatha, a correspondent in Siliconindia magazine, Sundarram discussed India's appeal to multinational manufacturers, the impact of Industry 4.0 on global and Indian manufacturing, and addressed data security in connected multinational manufacturing environments.
Why is India enticing for multinational manufacturers, and how does Industry 4.0 reshape global and Indian manufacturing?
It appears obvious to most multinationals that because of low-cost, English-speaking high skilled labour (including engineering skills) and the vast market India is, that it’s a very attractive place to be. But this is not the complete story. For instance, a multinational may think it could simply import products made overseas and sell them here. The fallacy in this is that the Indian customer’s needs are different not just from a cost perspective, but also factors like robustness to withstand harsh conditions and usage . Also, the Government’s “Make in India” push makes it increasingly unattractive to import (I’d rather not go into specifics here) and increasingly attractive to make (not just PLI, but other things as well).
What mechanization and standardization did in Industry 1.0, what electrification did in Industry 2.0, and what PLC’s and computerization did in Industry 3.0, the drivers of Industry 4.0 will do in this Industrial revolution. Virtual and augmented reality (VR & AR) as well as Additive manufacturing will impact the way manufacturing is done. But the really big changes will come from the Industrial Internet of Things (IIoT) and Artificial intelligence. The big change in every industrial revolution is productivity and IIoT with AI will I believe, be the game changer here (although VR/AR & Additive manufacturing will contribute). Many factories do not even measure OEE for instance, and those that do struggle to improve it. As an example, OEE could be as low as 40% and can be enhanced to over 80%, thereby doubling production with no capex addition. This is not a fantasy, but one of several real-life examples. Another case- OEM’s, particularly in the automobile industry, are able to track output of distant vendors to ensure they do not suffer supply disruption. Equipment suppliers can monitor equipment health remotely and even offer customers “Equipment as a service”, monitoring and charging for actual usage. These are examples where productivity improves by orders of magnitude, not just through incremental improvements, but even by a change of the business model.
How would we prioritize and strategize the implementation of IIoT solutions in a diverse and geographically spread manufacturing setup?
Top management could delegate (abdicate) this to the IT function is going to prioritize and strategize IIoT implementation, but they would be derelict in doing so. However, the reason for choosing this easy option is that Top management by and large, has little understanding of what IIoT is (and isn’t) or what it’s benefits are and more importantly, what are the strategic choices that are theirs and theirs alone to make.
The first question for the multi plant scenario you envisage is whether we need a common system across plants, or even whether you need a common system for all areas of a single plant. This then determines who this decision would be delegated to. For instance a plant head may get an “offer” from an equipment supplier for a “free” IoT system that “comes with the machine”, but which then puts the company in a systemic straitjacket. So this decision needs to be with a team representing both the plant and IT. Top management also has to decide on whether it wants a cloud based system or an edge based system or a hybrid, based on the pros and cons, with inputs from IT. Here it also has to decide on what kind of AI tools it wants to use (based on business considerations, not IT recommendations) as this may have an impact on whether the cloud is needed or not.
A common error that happens is confusing IIoT with Manufacturing Execution system (MES) or the Industrial control system (ICS) and believing that all this goes together and “everything must be connected”. But it is a decision as to whether MES and ICS are a part of the IIoT system, given the security implications, and if so whether ERP also is a part of the Grand design.
Having decided on the above foundational requirements, Top management has to decide whether to do this in a modular way or simultaneously across the plant (or even across plants). Managements may err here based on their ERP experience, but unlike ERP systems, IIoT can be implemented in modular fashion. Assuming they follow the “normal” strategy of modular implementation, the choice would be which machines in which shop in which plant to start with, and for which functions- Production or Quality or Safety or any other? Where is the highest return, that would motivate the organization to move forward? Then would come the decision to create an implementation team having stakeholders, operational technology (OT) experts and IT experts and firing them up to get on with the task.
The biggest challenge would be in how Top management brings in cultural change alongside modular implementation, which will be the make or break for IIoT in the organization. And management needs to decide on security options.
Considering data security concerns, how would we ensure the protection of sensitive information in a connected manufacturing environment, especially in the context of multi- national operations?
Security means different things for the IT environment and for the OT environment. In the IT context, it is about protecting data. In the OT context it is about protecting people and equipment, not data. Take an example data set generated in the production function of an IIoT system- “Machine no… was down for hours from to on In shift no due to “is of no value to a hacker. However, if the hacker can access and control the machine’s PLC and drives, he can do untold damage. This is why the ICS needs to be secure and as I commented earlier, management needs to decide how important it is to integrate the ICS into the IIoT environment. While hacking the MES may not enable the hacker to damage equipment or harm people, he can still bring the production system to a standstill. Strong backup plans for ICS & MES need to be available in case of a cybersecurity incident. Additionally, the MES may have valuable data about recipes, manufacturing sequences, SOP’s etc. and therefore needs better protection.
Another problem arises if there is connectivity between the ERP system where data confidentiality is paramount and the OT system where control is paramount. Either one could be used as a “back door” for entering the other. One possible solution is to have completely different logins, different ID’s and passwords, but the network segmentation approach needs to be decided by IT and presented to management.
What strategies would we need to employ to upskill and train the existing workforce to adapt to the technological shifts brought about by Industry 4.0 in the manufacturing sector?
If we take IIoT, , at the operator level, the training required would be minimal, sensitization is really what will be required. What I mean by this is that their work will actually be easier as they will have to make fewer manual entries. However they may feel a sense of “big brother is watching you”, when they realize that every event is being recorded and will come under scrutiny. At the supervisory and management level, a great deal of training is required to be able to use the system effectively. This means going beyond identifying downtime reasons for instance, and going into their causes, and then into the causes for those causes to ultimately arrive at actions for improvement. The increasing use of AI tools here will pose a learning challenge for middle management. What I visualize is a chatbot which analyzes factory performance and emerges with potential solutions. While this may appear simple, a tremendous amount of learning will be required for people to be able to query effectively as also to understand which tools to use where. For instance, Generative AI touted as a game changer in many areas may serve only a limited purpose in IIoT, where precise answers are required and the questions are close-ended, if not binary. Rule based algorithms and Machine Learning will be the weapons of choice in understanding shop floor problems and arriving at their solutions. Culturally too, factory management will have to be trained to be sensitive and not use the data generated like “big brother”.
Augmented reality tools are a great support for operators, particularly in complex operations, and while they may need some training, it is more a case of getting used to them and once done they will greatly aid productivity and mistake proofing. Virtual reality will increase the pace, improve safety and lower the cost of training workmen.
From a strategic standpoint, how would we position a multinational manufacturing company as an industry leader in the era of Industry 4.0, both in India and globally?
First and foremost is the product strategy ie what products the company is planning in India and how they are positioned. Some companies may choose the policy of “what’s good for USA is good for India” and slightly more enlightened ones would follow a policy of “throw spaghetti against the wall and see what sticks”. I would strongly recommend a more planned approach of researching, followed by test marketing. This needs to be followed by a decision of what to make and what to import. This could be as simple as -make low priced, low margin, high volume products, import high priced, high margin, low volume products. Sometimes however it’s not quite so simple. Other considerations like what is cheap and easy to make in India (and what is not) also come into play.
Coming to the IIoT aspect, multinational companies should realize that India is the best testing and development ground for IIoT and here is where they need to develop a system that can be replicated globally. IIoT development requires a multiplicity of skills- IT, OT, and Manufacturing management. Putting a team like this together and building an IoT system through experimentation, is an extremely labour-intensive process that will be very expensive in Europe or USA(given the cost of these skills), very difficult in places like Vietnam or Malaysia(given the availability of these skills) and undesirable in places like Russia or China. The process is labour intensive because there is no “standard” or “off the shelf” system available. Every manufacturing process is different, factories are different even for the same processes, and machines and layouts are different for similar factories. Why are they so? It usually wasn’t planned that way, it just happened. Using India as a testing center not just for IR 4.0 initiatives, but for all other (high skill) labour intensive initiatives should be embedded in the strategic thinking. One example would be Global Manufacturing systems- a function that seeks to improve manufacturing across factories globally, and goes hand in glove with IIoT development.
Note that I said “improve” and not “standardize” in the above statement as this brings me to the “Great standardization” trap that multinationals may fall into, believing that processes and policies developed in USA or Europe can be replicated in India. A simple real life example- A company may standardize a particular brand of drives for all machines across the globe, but in a particular geography support from that company is weak, and another brand is the market leader there. It would not be smart to follow the policy there. As technology gets more complex, there would be a temptation to try and shoehorn every factory’s systems and processes into a global policy. However, every geography is unique (and India has several geographies, each of which is unique), and every factory is unique. Failure to understand this is the most common cause of failure of a multinational in manufacturing in a particular geography.
"If there is connectivity between the ERP system where data confidentiality is paramount and the OT system where control is paramount. Either one could be used as a “back door” for entering the other"
Considering the potential for disruption, how would you develop contingency plans to mitigate risks associated with the implementation of new technologies in manufacturing plants?
Balancing risk and reward. Some would say that we can derive the full benefits only if we are fully connected and cloud based. I’d say that even were that true, I’d be satisfied with 80% of the gains with 20% of the risk. So I’d aim to minimize risk rather than manage it. As I had mentioned earlier, implementation should not be all at once but in modular fashion, one factory, one shop and a few machines, starting preferably with downtime monitoring which can capture low hanging fruit. Top management, familiar with ERP implementations may not be able to grasp this possibility and may be in an “all or nothing” mindset. As implementation expands, new learning is incorporated into further implementation. You may find for example, that a particular sensor does not work well in your environment, and you may have to change the brand or maybe even the measurement methodology. Or you may find that your system integrated well with some Profinet based PLC’s but don’t work well in the next stage with older machines with Modbus based PLC’s.
Secondly, as I mentioned earlier Top management should not be swayed by those who want to connect everything, after being told that the system is “absolutely hackproof”. There is no such thing, everything can be hacked given expertise, time and money. Keep the ERP & MES separate from the IIoT, preferably with different user names and passwords. Use active cyber defense and do not expose the ICS to the internet unless absolutely necessary.
Finally prepare for the worst-case scenario, unlike people in Japan who did not have paper money after the recent earthquakes and couldn’t buy food and other essentials as the internet was down. Paper systems, looked down on in today’s tech environment are a foolproof backup.
In a conversation with Charulatha, a correspondent in Siliconindia magazine, Sundarram discussed India's appeal to multinational manufacturers, the impact of Industry 4.0 on global and Indian manufacturing, and addressed data security in connected multinational manufacturing environments.
The big change in every industrial revolution is productivity and IIoT with AI will I believe, be the game changer here (although VR/AR & Additive manufacturing will contribute)
Why is India enticing for multinational manufacturers, and how does Industry 4.0 reshape global and Indian manufacturing?
It appears obvious to most multinationals that because of low-cost, English-speaking high skilled labour (including engineering skills) and the vast market India is, that it’s a very attractive place to be. But this is not the complete story. For instance, a multinational may think it could simply import products made overseas and sell them here. The fallacy in this is that the Indian customer’s needs are different not just from a cost perspective, but also factors like robustness to withstand harsh conditions and usage . Also, the Government’s “Make in India” push makes it increasingly unattractive to import (I’d rather not go into specifics here) and increasingly attractive to make (not just PLI, but other things as well).
What mechanization and standardization did in Industry 1.0, what electrification did in Industry 2.0, and what PLC’s and computerization did in Industry 3.0, the drivers of Industry 4.0 will do in this Industrial revolution. Virtual and augmented reality (VR & AR) as well as Additive manufacturing will impact the way manufacturing is done. But the really big changes will come from the Industrial Internet of Things (IIoT) and Artificial intelligence. The big change in every industrial revolution is productivity and IIoT with AI will I believe, be the game changer here (although VR/AR & Additive manufacturing will contribute). Many factories do not even measure OEE for instance, and those that do struggle to improve it. As an example, OEE could be as low as 40% and can be enhanced to over 80%, thereby doubling production with no capex addition. This is not a fantasy, but one of several real-life examples. Another case- OEM’s, particularly in the automobile industry, are able to track output of distant vendors to ensure they do not suffer supply disruption. Equipment suppliers can monitor equipment health remotely and even offer customers “Equipment as a service”, monitoring and charging for actual usage. These are examples where productivity improves by orders of magnitude, not just through incremental improvements, but even by a change of the business model.
How would we prioritize and strategize the implementation of IIoT solutions in a diverse and geographically spread manufacturing setup?
Top management could delegate (abdicate) this to the IT function is going to prioritize and strategize IIoT implementation, but they would be derelict in doing so. However, the reason for choosing this easy option is that Top management by and large, has little understanding of what IIoT is (and isn’t) or what it’s benefits are and more importantly, what are the strategic choices that are theirs and theirs alone to make.
The first question for the multi plant scenario you envisage is whether we need a common system across plants, or even whether you need a common system for all areas of a single plant. This then determines who this decision would be delegated to. For instance a plant head may get an “offer” from an equipment supplier for a “free” IoT system that “comes with the machine”, but which then puts the company in a systemic straitjacket. So this decision needs to be with a team representing both the plant and IT. Top management also has to decide on whether it wants a cloud based system or an edge based system or a hybrid, based on the pros and cons, with inputs from IT. Here it also has to decide on what kind of AI tools it wants to use (based on business considerations, not IT recommendations) as this may have an impact on whether the cloud is needed or not.
A common error that happens is confusing IIoT with Manufacturing Execution system (MES) or the Industrial control system (ICS) and believing that all this goes together and “everything must be connected”. But it is a decision as to whether MES and ICS are a part of the IIoT system, given the security implications, and if so whether ERP also is a part of the Grand design.
Having decided on the above foundational requirements, Top management has to decide whether to do this in a modular way or simultaneously across the plant (or even across plants). Managements may err here based on their ERP experience, but unlike ERP systems, IIoT can be implemented in modular fashion. Assuming they follow the “normal” strategy of modular implementation, the choice would be which machines in which shop in which plant to start with, and for which functions- Production or Quality or Safety or any other? Where is the highest return, that would motivate the organization to move forward? Then would come the decision to create an implementation team having stakeholders, operational technology (OT) experts and IT experts and firing them up to get on with the task.
The biggest challenge would be in how Top management brings in cultural change alongside modular implementation, which will be the make or break for IIoT in the organization. And management needs to decide on security options.
Considering data security concerns, how would we ensure the protection of sensitive information in a connected manufacturing environment, especially in the context of multi- national operations?
Security means different things for the IT environment and for the OT environment. In the IT context, it is about protecting data. In the OT context it is about protecting people and equipment, not data. Take an example data set generated in the production function of an IIoT system- “Machine no… was down for hours from to on In shift no due to “is of no value to a hacker. However, if the hacker can access and control the machine’s PLC and drives, he can do untold damage. This is why the ICS needs to be secure and as I commented earlier, management needs to decide how important it is to integrate the ICS into the IIoT environment. While hacking the MES may not enable the hacker to damage equipment or harm people, he can still bring the production system to a standstill. Strong backup plans for ICS & MES need to be available in case of a cybersecurity incident. Additionally, the MES may have valuable data about recipes, manufacturing sequences, SOP’s etc. and therefore needs better protection.
Another problem arises if there is connectivity between the ERP system where data confidentiality is paramount and the OT system where control is paramount. Either one could be used as a “back door” for entering the other. One possible solution is to have completely different logins, different ID’s and passwords, but the network segmentation approach needs to be decided by IT and presented to management.
What strategies would we need to employ to upskill and train the existing workforce to adapt to the technological shifts brought about by Industry 4.0 in the manufacturing sector?
If we take IIoT, , at the operator level, the training required would be minimal, sensitization is really what will be required. What I mean by this is that their work will actually be easier as they will have to make fewer manual entries. However they may feel a sense of “big brother is watching you”, when they realize that every event is being recorded and will come under scrutiny. At the supervisory and management level, a great deal of training is required to be able to use the system effectively. This means going beyond identifying downtime reasons for instance, and going into their causes, and then into the causes for those causes to ultimately arrive at actions for improvement. The increasing use of AI tools here will pose a learning challenge for middle management. What I visualize is a chatbot which analyzes factory performance and emerges with potential solutions. While this may appear simple, a tremendous amount of learning will be required for people to be able to query effectively as also to understand which tools to use where. For instance, Generative AI touted as a game changer in many areas may serve only a limited purpose in IIoT, where precise answers are required and the questions are close-ended, if not binary. Rule based algorithms and Machine Learning will be the weapons of choice in understanding shop floor problems and arriving at their solutions. Culturally too, factory management will have to be trained to be sensitive and not use the data generated like “big brother”.
Augmented reality tools are a great support for operators, particularly in complex operations, and while they may need some training, it is more a case of getting used to them and once done they will greatly aid productivity and mistake proofing. Virtual reality will increase the pace, improve safety and lower the cost of training workmen.
From a strategic standpoint, how would we position a multinational manufacturing company as an industry leader in the era of Industry 4.0, both in India and globally?
First and foremost is the product strategy ie what products the company is planning in India and how they are positioned. Some companies may choose the policy of “what’s good for USA is good for India” and slightly more enlightened ones would follow a policy of “throw spaghetti against the wall and see what sticks”. I would strongly recommend a more planned approach of researching, followed by test marketing. This needs to be followed by a decision of what to make and what to import. This could be as simple as -make low priced, low margin, high volume products, import high priced, high margin, low volume products. Sometimes however it’s not quite so simple. Other considerations like what is cheap and easy to make in India (and what is not) also come into play.
Coming to the IIoT aspect, multinational companies should realize that India is the best testing and development ground for IIoT and here is where they need to develop a system that can be replicated globally. IIoT development requires a multiplicity of skills- IT, OT, and Manufacturing management. Putting a team like this together and building an IoT system through experimentation, is an extremely labour-intensive process that will be very expensive in Europe or USA(given the cost of these skills), very difficult in places like Vietnam or Malaysia(given the availability of these skills) and undesirable in places like Russia or China. The process is labour intensive because there is no “standard” or “off the shelf” system available. Every manufacturing process is different, factories are different even for the same processes, and machines and layouts are different for similar factories. Why are they so? It usually wasn’t planned that way, it just happened. Using India as a testing center not just for IR 4.0 initiatives, but for all other (high skill) labour intensive initiatives should be embedded in the strategic thinking. One example would be Global Manufacturing systems- a function that seeks to improve manufacturing across factories globally, and goes hand in glove with IIoT development.
Note that I said “improve” and not “standardize” in the above statement as this brings me to the “Great standardization” trap that multinationals may fall into, believing that processes and policies developed in USA or Europe can be replicated in India. A simple real life example- A company may standardize a particular brand of drives for all machines across the globe, but in a particular geography support from that company is weak, and another brand is the market leader there. It would not be smart to follow the policy there. As technology gets more complex, there would be a temptation to try and shoehorn every factory’s systems and processes into a global policy. However, every geography is unique (and India has several geographies, each of which is unique), and every factory is unique. Failure to understand this is the most common cause of failure of a multinational in manufacturing in a particular geography.
"If there is connectivity between the ERP system where data confidentiality is paramount and the OT system where control is paramount. Either one could be used as a “back door” for entering the other"
Considering the potential for disruption, how would you develop contingency plans to mitigate risks associated with the implementation of new technologies in manufacturing plants?
Balancing risk and reward. Some would say that we can derive the full benefits only if we are fully connected and cloud based. I’d say that even were that true, I’d be satisfied with 80% of the gains with 20% of the risk. So I’d aim to minimize risk rather than manage it. As I had mentioned earlier, implementation should not be all at once but in modular fashion, one factory, one shop and a few machines, starting preferably with downtime monitoring which can capture low hanging fruit. Top management, familiar with ERP implementations may not be able to grasp this possibility and may be in an “all or nothing” mindset. As implementation expands, new learning is incorporated into further implementation. You may find for example, that a particular sensor does not work well in your environment, and you may have to change the brand or maybe even the measurement methodology. Or you may find that your system integrated well with some Profinet based PLC’s but don’t work well in the next stage with older machines with Modbus based PLC’s.
Secondly, as I mentioned earlier Top management should not be swayed by those who want to connect everything, after being told that the system is “absolutely hackproof”. There is no such thing, everything can be hacked given expertise, time and money. Keep the ERP & MES separate from the IIoT, preferably with different user names and passwords. Use active cyber defense and do not expose the ICS to the internet unless absolutely necessary.
Finally prepare for the worst-case scenario, unlike people in Japan who did not have paper money after the recent earthquakes and couldn’t buy food and other essentials as the internet was down. Paper systems, looked down on in today’s tech environment are a foolproof backup.
.jpg "Pixel Web Series: MVPS That Act As Catalyst Of Growth")
