276
D. Mishra et al.
8.3.4.2 Real-Time Capability
Once the communication is established between the machine and the
computer/computing devices, the next thing to ensure is the real-time flow of the
information. Here, information refers to the data which is acquired from the sensors
engaged in the manufacturing process, both for process and machine health monitoring. “Delay” or lag in the data transmission is a concern in monitoring as it
will assess the timeliness of the real-time system. This is followed by the real-time
analysis of the data collected for finding useful inferences about the manufacturing
process. This could be directly realized for the welding techniques. A recent article
mentions the Industry 4.0 attributes in GTAW technique and refers to the utilization
of a software termed, “Open Platform Communications Unified Architecture” [92].
This software is helpful for machine-to-machine communication and can be configured with a cloud server for data transmission. The software recognizes a machine
from its Internet protocol (IP) address.
8.3.4.3 Decentralization
This principle refers to the decision-making in real time for manufacturing processes
[93]. It aims at making the machines operate in an uninterrupted manner, also without
any human intervention. In order to fulfil this, the machine needs to be embedded
with several sensors. The reason of integrating several sensors has already been
discussed. Once “interoperability” and “real-time capability” have been ensured, the
exercise of extracting meaningful information comes into picture. This would require
various open-source software for data analytics, followed by ML, DL and AI. The
use of open-source platforms will provide: (a) flexibility in solving several problems,
(b) agility for solving the problems in multiple ways, (c) speed by listing a range
of services to choose from, (d) cost-effective solution and (e) easy use across an
organization since it is available openly. This principle will help in developing an
automated environment where the decisions are being taken by the system without
any human involvement in real time. In reference to welding, the ML/DL can be
utilized for predicting the weld quality, identifying the welding defects and faults in
the power source/welding machine in real time. This will be followed by the corrective actions. For implementation of ML/DL, a “knowledge base” will be required.
This knowledge base will consist of the data belonging to the production machine,
acquired information from sensors, and environmental conditions prevailing during
welding. This knowledge base needs to be enriched with each passing day so that
the learning of the machine gets strengthened.
D. Mishra et al.
8.3.4.2 Real-Time Capability
Once the communication is established between the machine and the
computer/computing devices, the next thing to ensure is the real-time flow of the
information. Here, information refers to the data which is acquired from the sensors
engaged in the manufacturing process, both for process and machine health monitoring. “Delay” or lag in the data transmission is a concern in monitoring as it
will assess the timeliness of the real-time system. This is followed by the real-time
analysis of the data collected for finding useful inferences about the manufacturing
process. This could be directly realized for the welding techniques. A recent article
mentions the Industry 4.0 attributes in GTAW technique and refers to the utilization
of a software termed, “Open Platform Communications Unified Architecture” [92].
This software is helpful for machine-to-machine communication and can be configured with a cloud server for data transmission. The software recognizes a machine
from its Internet protocol (IP) address.
8.3.4.3 Decentralization
This principle refers to the decision-making in real time for manufacturing processes
[93]. It aims at making the machines operate in an uninterrupted manner, also without
any human intervention. In order to fulfil this, the machine needs to be embedded
with several sensors. The reason of integrating several sensors has already been
discussed. Once “interoperability” and “real-time capability” have been ensured, the
exercise of extracting meaningful information comes into picture. This would require
various open-source software for data analytics, followed by ML, DL and AI. The
use of open-source platforms will provide: (a) flexibility in solving several problems,
(b) agility for solving the problems in multiple ways, (c) speed by listing a range
of services to choose from, (d) cost-effective solution and (e) easy use across an
organization since it is available openly. This principle will help in developing an
automated environment where the decisions are being taken by the system without
any human involvement in real time. In reference to welding, the ML/DL can be
utilized for predicting the weld quality, identifying the welding defects and faults in
the power source/welding machine in real time. This will be followed by the corrective actions. For implementation of ML/DL, a “knowledge base” will be required.
This knowledge base will consist of the data belonging to the production machine,
acquired information from sensors, and environmental conditions prevailing during
welding. This knowledge base needs to be enriched with each passing day so that
the learning of the machine gets strengthened.