T Nallaperumal says....‘The manufacturer needs to upgrade his knowledge’
IIoT is said to have the potential to transform every aspect of manufacturing. What is the vendors’ perspective here?
All the vendors are getting ready with the latest technologies of IIoT, AI, ML, etc., else, they will be out of business within the next one year.
The topic going on the internet is “How Alexa will effectively do the parenting!” If the OEMs/companies don’t shift their work culture and augment their workforce, they can't sustain within the next 5 years. This Industrial Revolution 4.0 can be analogy with the computerisation during the early 90s. If the machines are not equipped with the IoT knowledge the intelligence of the machine is wasted and of course it leads to productivity loss and the ecosystem will fall down. Not only the vendor needs to upgrade his knowledge, but also the manufacturer also needs to upgrade his knowledge.
Digital transformation is the key to success in IIoT. Are the user industries ready for this?
After the recent term of Industry 4.0, all the consumers want to catch the flight and be 'there'. But, no one knows, what the immediate step is to start with. Data is the key. And along with it, the architect should know which data to skip, during the process. This whole term needs to be considered as Digital Transformation. I personally think, not all of them are ready, with solutions for the problems they face. They would need an experienced person with background knowledge of machine design and computer to finalise the need of the company, which will vary, according to the bottlenecks they face.
While the advantages of IIoT can be overwhelming, it is often the case that manufacturers are in a dilemma where to begin. Is this the reality?
As discussed earlier, the starting point is the problem, and yes, many people do not know where to start. The OEM should know the problems they face in day-to-day operations, and these need to be listed down in a chronological order. Each problem needs to be analysed with the machine parameter, of how it is linked with the actual problem. While 75% of the knowledge is with the manufacturers, the remaining 25% needs to be decided by the IoT team, of how efficiently, it can be taken out. Which parameters need to have the Artificial Intelligence and which model needs to be decided for the Machine Learning. If the customer already has some big-data, then a data scientist is needed.
Do vendors have enough use cases to convince a fence sitting manufacturing enterprise?
The use cases are phenomenal. There are lot of use-cases and lot of solutions that are available. The success of the implementation is that, how soon, it can done, and which ML (machine learning) needs to be used for that particular use case. This can be done only with an experienced solution provider, who has in-depth knowledge in the manufacturing industry for at least 10 years. Most of the MNCs have already implemented or are in the process of implementing Industry 4.0 in Europe. In the USA, the same is considered as Smart Factory. To name a few, Siemens has lot of use cases in the Mindsphere and Allen Bradley, ABB, etc., have also done lot of use cases with their customer base.
Ideally, for a typical user, legacy equipment is a big consideration as to how to start the digital transformation. Are there solutions for this?
The problem faced by the implementation is exactly the same. The OEMs/manufacturers have many islands of data, and they don't know how to take it forward from there. This is no single vendor who can completely combine all the data and take it up to a particular protocol that is needed. Hence, finding out the right partner – and cost effective – is a difficult task at the moment. We can see this in two different views, in a Greenfield project, it is much easier, as the manufacturer and the solution provider can have an elaborate discussion on how to take the data with different protocols, like OPC-UA, etc., and accordingly the machine supplier needs to deliver the machine. The cloud part shall be handled efficiently, based on the new technologies and data available. Alternatively, on a Brownfield project, getting the required data in a particular protocol is very difficult, based on the machine life and the technologies used inside the panel. Hence, new sensors/upgradation of the machine needs to be done, to make it IoT enabled, and this is a costly affair, unless the consultant knows the macro level understanding of the machine, PLCs, sensors, protocol, cloud, model creation, AI and ML, delivery options, etc.
Is it possible to start small, and then scale up the implementation?
It is logically and practically possible to start with a small step for Phase 1. And then the possibility of extending throughout the factory needs to be analysed. If the full effort is put initially, then the expectation and the output shall not match, and obviously the costs will increase, which might lead to failure in many cases. The role of the top management is also needed, to fix the needs for each phase and work out in a gradual and steady phase.
What will be the next step of IIoT?
Digitisation of the factory is one part. Moving to the reality, in first case, the machines need to communicate between each other and share the 'failures' to the other machines. For example, in a CNC machine, a particular part number was running, and a tool has failed, then it should be communicated with the next machine that is processing the subsequent part number, to reduce the speed or some other variant, so that the tool doesn't get broken again.
The second case, the IIoT values should be used for preventive maintenance, incorporating the Machine Learning and AI together. The machine should have the intelligence to predict, whether the machine is going to fail in the near future, based on the analysis of the previous data.
Appville softwares is a technology driven company, focusing on Industry 4.0 solutions, including Augmented Reality, IIoT and Digital Twin. They use AI and ML for their predictive analysis module.