Industry 4.0, Convergence of IOT and AI
Industry 4.0 is the name given to a combination of cyber-physical systems (CPS), the Internet of things (IOT), cloud computing and cognitive computing. It is also called the Fourth Industrial Revolution. It comprises of an intelligent network of machines and the manufacturing system. It aims at making the manufacturing process more efficient and automated. It is data-driven.
Industry 4.0 (I4) was first introduced as a project of the German Government, the aim of which was to transform industrial manufacturing through digitalization. Gradually, it became synonymous with Industrial IOT and came to be known widely.
Technologies enabling Industry 4.0
- Cloud Computing
- Big Data
- Artificial Intelligence
- Additive manufacturing
- Augmented Reality/Virtual Reality (AR/VR)
Industrial IOT (IIOT)
The Industrial IOT is a key element of Industry 4.0. It incorporates machine learning and big data technologies to harness the data from sensors, machine-to-machine communication and automation. Smart machines are accurate and consistent in capturing and communicating real-time data. This data enables companies to pick up on inefficiencies and problems sooner, saving time and money and supporting business intelligence efforts.
Three-Tier IIOT Architecture
Data flows and control flows take place through the three-tier architecture pattern connected through networks.
1. Edge Tier
Edge Tier is where there are devices, e.g. In a manufacturing system or warehouse. Edge nodes collect data and push data to the gateway. The edge gateway collects and processes things in real-time. There is data acquisition, data processing and some decision making, device management and security.
2. Platform Tier
Platform Tier consists of components for IIOT system. It consists of fundamental services to run applications. It consolidates data, routes commands between the edge and domain applications; and provides a platform for domain applications, security, device management. The connection is typically via cloud.
3. Enterprise Tier
The enterprise architecture runs on the cloud where the domain applications run. It comprises of two parts- first, the IIOT domain application, and second, the IIOT application with which it will interact. The IIOT domain application provides a user interface. The data is consumed here. The IIOT application with which it interacts may be something like supply chain management (SCM) or a product system. This creates a smart manufacturing system.
Smart manufacturing is also known as digital transformation. Why do we need digital transformation?
To overcome challenges such as-
- Faster time to market
- Real-time visibility
- Improve asset utilization
- Improve workforce productivity and safety
- Improve product quality — reduce rejections
- Reduce downtime
- Just-in-time maintenance
Application areas within smart manufacturing include creating a digital twin. A digital twin is a digital replica of a physical asset, used to optimally manage that asset. It uses technologies such as Big Data/AI/ML. It is maintained throughout the lifecycle of the asset i.e. from the design data of the machine, to the operations data, maintenance data, and finally the ‘end of life’ data. There is a constant collaboration with R&D to send back the production data to the design stage and thereby make any improvements. Digital modeling and simulations such as 3D printing are used to simulate the scenario while designing the product, to ensure it handles the scenarios well.
Smart Manufacturing thereby improves performance and quality of the process and product, predicts problems, and prevents downtime.
End-to-end visibility can be implemented using the digital supply chain. Suppliers are automatically informed if any items are required to be added. Without much human action, inventory management can be effectively implemented using IOT.
Data collected in real-time can be sent to the manufacturer. Manufacturers can plan and deliver parts efficiently based on the data shared by the system.
Multi-channel customer support can be possible. Manufacturers can also optimize their own internal support structure.
Using smart operations, one can get real-time Overall Equipment Effectiveness (OEE), condition monitoring of equipment, actionable insights and analytics. These help improve productivity, reduce downtime and improve quality.
Maintenance can be (a)Running hours based and (b) Prescriptive.
Huge cost saving is possible with just-in-time maintenance. The idle time of machine due to breakdown can be prevented with prescriptive maintenance. In addition to asset maintenance, tool maintenance is done too.
Tracking of the whole journey is possible in smart operations. It helps to monitor vendor performance. It also offers traceability to customers.
Usage of resources such as electricity, steam, compressed air need to be optimized. Also, long term analysis of demand and supply will make the system optimal.
Smart services means products that are converted into services. In smart services, the collected data of the devices in a network, is converted into valuable information through which the service provider sets up a new service, beneficial to the customer as well as to the company providing the service.
Machine Learning (ML) in Manufacturing
The supplier can easily manage an asset because of remote monitoring. IOT asset management allows for predictive maintenance that stops problems before they start. Unlike preventative maintenance, that is scheduled in advance at certain intervals, predictive maintenance uses performance data to determine if and when an asset is likely to fail, thus predicting maintenance needs.
ML methods used:
Predictive maintenance- Anomaly detection, Clustering
Failure prediction- Classification, Clustering
Critical components failure- Vibrational analysis
Downtime analysis- Clustering
ML provides meaningful insights through data about product items’ location, statuses, movements, etc., and giving users a corresponding output. This output could mean making changes in the production cycle or making design improvements. This could be integrated with other departments to bring about further meaningful insights. As a result, manufacturers can balance the amount of on-hand inventory, increase the utilization of machines, reduce lead time, and control and improve the quality.
ML methods used:
Quality control- Anomaly detection, Clustering
Utilities and Energy contribute to 30–35% of the cost. Optimal usage of utilities and resources such as electricity, steam, compressed air — use machine learning technologies. Besides these, fine-tuning, load balancing, energy saving, are other reasons to use combine ML with IOT technologies, for infrastructure management.
ML methods used:
Energy optimization- Regression
Originally published at https://thakkarnidhi.com on March 7, 2019.