Application of Internet of Things (IoT) and Artificial Intelligence (AI) in Welding Industry 4.0

Dr. Tushar Sonar, Doctoral Researcher (Project Associate), Centre for Materials Joining and Research (CEMAJOR), Department of Manufacturing Engineering, Annamalai University, Chidambaram, Tamil Nadu State, India.
Currently working on the ISRO Project in InterPulsed TIG Welding of Aerospace alloys at Annamalai University.

1.0 Introduction

1.1 What is Industry 4.0?

Industry 4.0 describes the growing trend towards automation and data exchange in technology and processes within the manufacturing industry, including: The internet of things (IoT), Cyber-physical systems (CPS), Smart manufacture, Smart factories, Cloud computing, Cognitive computing, Artificial intelligence. The first industrial revolution came with the advent of mechanisation, steam power and water power. This was followed by the second industrial revolution, which revolved around mass production and assembly lines using electricity. The third industrial revolution came with electronics, I.T. systems and automation, which led to the fourth industrial revolution that is associated with cyber physical systems (Figure 1).

1.2 What is the Internet of Things (IoT) and Artificial Intelligence (AI)?

The IoT essentially refers to the idea of connecting anything that is powered to both the internet and each other. This includes everything from mobile phones, kitchen appliances and office equipment, right through to jet engines and large-scale manufacturing equipment. The move towards wide-scale uptake of IoT had already started, fuelled by improved Broadband Internet availability, lower connection and technology costs, increased smartphone usage, and the fact that more and more devices, appliances and machines are being built with Wi-Fi capability.

Artificial intelligence (AI) is the simulation of human intelligence processes by machines, especially computer systems. These processes include learning (the acquisition of information and rules for using the information), reasoning (using rules to reach approximate or definite conclusions) and self-correction. AI is incorporated into different types of technology.

• Automation: What makes a system or process function automatically. For example, robotic process automation(RPA) can be programmed to perform high-volume, repeatable tasks that human normally performed.
• Machine learning:The science of getting a computer to act without programming.
• Machine vision:This technology captures and analyses visual information using a camera, analog-to-digital conversion and digital signal processing.
• Robotics:A field of engineering focused on the design and manufacturing of robots. Robots are used to do tasks that are difficult for humans to perform or perform consistently.

Fig.1 Evolution of Industry 4.0

The main attractions of the IoT and AI are:

• The “user” of the devices can obtain useful information that they can use to improve their specific circumstances. The information would typically be based on data collected by their own smart devices which has been analysed and interpreted by remote service providers.
• The broader industry and society can use aggregated data from thousands or millions of devices to analyse problems and trends. This information can be used for anything from regulatory and research purposes to marketing purposes and many other purposes in between.

There are three key uses for the IoT and AI within the manufacturing industry: –

1. Manufacturing operations
2. Production asset management and maintenance
3. Field service

In manufacturing operation, IoT and AI is being used for purposes such as intelligent manufacturing, monitoring and enhancing performance, human-machine interaction, operational visibility and production line planning. It is estimated that 57% of IoT spending within the global manufacturing industry is related to manufacturing operations.

IoT and AI is used in production asset maintenance and management to monitor key parameters related to performance, damage and breakdowns, bottlenecks, quality and a range of other factors. This means that the IoT is being used not only to improve performance, but also preventative maintenance.

Manufacturers are also using IoT and AI when they take on the role of service providers. This includes product and business-related services, which are both vital in providing growth to manufacturers and the wider-industry. The ability to move information through digital networks and an IoT-enabled manufacturing ecosystem is crucial in improving customer-service delivery.

2.0 How Can the IoT and AI impact Welding?

The IoT and AI has the potential to improve almost every aspect of a welding workshop. It can store the latest welding procedure regulations, manage and renew qualifications of welders, provide improved quality control, verify product quality, detect and place orders for consumables and gas, suggest training requirements, and provide welding project management assistance.

2.1 The Benefits of Introducing the IoT and AI to a Welding Workshop

One of the biggest problem welding industries in developed countries facing is a skills shortage. In the US, for example, the American Welding Society predicts a shortage of 400,000 welding operators by 2024, with the supply-demand gap a consequence of an ageing workforce (average age of a welder is 57).The age profile of workers within the welding industry has been skewed to such an extent that crippling skills shortages of highly skilled welders is now inevitable. We’ve now moved well past the point of potentially averting this crisis through skills attraction. The question we now collectively face is how to respond constructively within the context of this unavoidable challenge.

The IoT and AI can alleviate this problem in several ways. An IoT-enabled welding process does not rely on a skilled welder as much as a manual process. An unskilled, relatively untrained worker can complete the same task as a highly-skilled, experienced welder if they are working on a machine that is fitted with IoT-enabled artificial intelligence.

Both these concepts are tied to the idea of automation. Automation has been part of the factory floor for decades now, but IoT ushers in a form of automation that is very different from the clunky robot performing tedious and simplistic tasks on the assembly line.

The idea of Industry 4.0, which is closely tied to IoT, “represents the vision of the interconnected factory where equipment is online, and in some way is also intelligent and capable of making its own decisions.”

In a welding workshop that is plugged into the world of the IoT, the automatic welding process becomes dynamic and capable of responding to parameters, the workpiece and other external factors. Essentially, the welding process becomes a thinking entity, capable of reacting to change in the way a human welder would. Not only does this address a chronic skills shortage, but it also makes for a more productive workplace. Artificial Intelligence and machine learning allow for projects to be completed faster, without sacrificing quality.

If humans are forced to work faster, quality inevitably suffers, and the need for re-work all but cancels out any improvements in production. This isn’t the case with automated welding procedures that are capable of learning via data input. The IoT is the future of fabrication, and in many ways, it’s the future of the world as we know it. Everything will be connected, so a failure to invest and learn how to create a connected workplace is a failure to stay in touch with the direction of the industry, and the consumer base that workshops service.

3.0 IoT and Artificial Intelligence in current welding industry

Using the Controller (Figure 2), you can connect to K-TIG’s expert support team with just the touch of a button. K-TIG’s support engineers can review your system, work with you to resolve any issues and ensure you’re up and running in record time. You can even opt in for a system Health Check, during which K-TIG’s expert engineers review your system for peak performance and maintenance requirements.

Fig.2 K-TIG Controller

The Controller’s WiFi and ethernet connectivity allow you to upload and share weld procedures, and review and store comprehensive weld reports. Equipped with easily accessible dual powered USB ports, operators can even use the Controller to charge other devices such as phones and tablets.

4.0 How the IoT and AI Could Impact the World of the Welder?

Many automated welding machines are already connected to a computer in some way or another. Most modern welding power sources have computer-based control systems. Many of these power sources can be interfaced with a network. Many power sources are in fact networked and could be accessed at any time from a computer anywhere in the world. In this regard, the IoT has already arrived in the world of automated and robotic welding. The first obvious application would be in diagnostics and calibration of the equipment itself, as the equipment could be interfaced with a network on a routine basis to perform those functions. This application will however not have much impact on the job of the field Welder in a direct way.

Future IoT application to the field Welder will in all probability arise mainly as part of an artificial intelligence platform, as the artificial intelligence platform would be highly advantageous in performing control and monitoring functions even while the equipment is not connected to the internet. For continuous IoT functionality, continuous internet access would obviously be required. The artificial intelligence would make it much more effective to add all kinds of sensors to the Welder to record the necessary parameters and inputs. This would be the case whether the device is constantly connected to the internet or not. Machine learning vision systems would be able to potentially perform measurements such as welding travel speed, which is necessary for real time calculation of welding heat input, when combined with voltage and current inputs from the welding power source.

Another very useful sensor would be an infrared enabled vision system. While measuring welding heat inputs are widely used currently, they are not perfect measurements. The actual information that we want when measuring the heat input is the cooling rate of the weld and the heat affected zone (HAZ) of the base metal. With an infrared enabled vision system, the cooling rate calculation could be made directly, rather than relying on the imperfect heat input proxy that we currently rely on. With a continuous record of the weld cooling rate, we could have an almost fool-proof measure that Hydrogen Assisted Cold Cracking (HACC) has not occurred. This will mean that the need for time delays following welding, followed by ultrasonic testing, could be reduced or in many cases eliminated. In cases where high heat inputs are problematic, such as with the welding of stainless steels, this record could serve as assurance that the microstructures are not degraded, allowing the material’s full corrosion resistance to be realised.

5.0 What is Internet of Welding?

Internet of Welding must enable a company to make their operational processes leaner or speed up existing processes in order to reduce waste. Quality management, internal logistics and device maintenance are excellent examples of processes that can be improved with IoW. New welding machines know when they are running short of filler wire, when a weld joint is ready for inspection or when a job is finished.

The consistent implementation of Industry 4.0 also has a decisive influence on the factory of the future, which will develop into a smart factory. For a smart factory to work more or less autonomously without human interaction, some criteria must be met for welding applications:

• The welding know-how has to be digitized and prepared in such a way that a computer can make similarly correct decisions as an experienced welding technologist.
• The welding equipment must be equipped with high-performance information and communication technology and customized sensors so that all production-relevant information can be digitized and given the necessary real-time behaviour of the system.
• The need to transfer and store large amounts of data requires powerful network infrastructures and sufficient storage capacity.

5.1 Welding Manufacture Based on IOT and AI

With the technique development of Internet of Things (IOT), the production mode of the welding manufacturing industry has changed a lot. A brief framework of the smart factory based on IOT. In the framework, the smart factory consists of four layers: physical resource layer, industrial network layer, cloud layer, and supervision and control terminal layer. Physical resource layer is based on smart device that can communicate with each other through industrial network. Various information systems, such as manufacturing execution system (MES) and enterprise resource planning (ERP), exist in the cloud that can acquire massive data from the physical resource layer and interact with people through the terminals. This actually forms a cyber-physical system (CPS) where physical artifacts and informational entities are deeply integrated. A pyramid framework to demonstrate the architecture of intelligent factory based on IOT, from field devices (sensors/actuators) and programmable logic controllers (PLC) through process management and manufacturing execution systems (MES) to the enterprise level (ERP) software (Fig. 3).

Fig.3 A pyramid framework architecture of the intelligent factory base on IOT

5.2 Real-Time Welding Data to Optimize Quality, Efficiency

Welding data-management systems (Figure 4) gather data in real time from the shop floor, right from the welding machines (semiautomatic or robotic) to provide management with a live look at every welder and welding cell. Data can even move to the cloud (Internet or intranet) to allow for remote review over the Internet. Managers can keep a close watch over numerous productivity-related metrics, such as arc-on time and weld metal-deposition rate. Supplier-fabricators can use the systems to automate the preparation of documents they need to deliver to their customers, and alleviate welders from having to spend time logging data. And, large companies can centralize their welding-knowledge base. They can develop standard welding procedures based on application specifics, and push those procedures out to their weld cells or to the weld cells of their suppliers. Wondering how often welders (human or robotic) have to stop what they’re doing because of excess spatter or poor penetration, so they can evaluate and solve procedural issues? Welding-information-management systems will provide that knowledge as well, and allow shops to immediately identify and address common causes of downtime.

Operations managers use welding-information systems to seek opportunities to increase productivity while maintaining quality. They can identify which operators are the most productive and which may need additional training, measure true welding costs, discover opportunities to reduce costs and gauge the impact of continuous-improvement initiatives. Quality-assurance engineers can trace weld performance and determine what changes are needed to prevent future issues. And, they can find which welders are operating outside of acceptable limits and prevent future occurrences. Service and maintenance technicians who need to keep track of all of the machines in their fleet can instantly identify what each machine is doing. If a machine experiences hiccups, the information-management system can alert technicians so that they can provide quick and accurate diagnoses. And, welding engineers can use these systems to adjust and optimize a welding process for a given application, test it, and then download the updated parameters to one or more machines in the shop.

Fig.4 Welding data-management system.

With welding data-management systems, managers can keep a close watch over numerous productivity-related metrics, such as arc-on time and weld metal-deposition rate. Miller Electric introduced its Insight Centre point version 9.0 at FABTECH, which include a Smart Part Tracking feature that automatically calculates the amount of weld metal required based on an operator-input weld symbol and fillet size.

5.3Welding-Process Insight from Miller Electric Mfg. Co.

Miller has developed its Welding Intelligence systems to collect data from several models of the firm’s power sources. The data can transmit via a wired Ethernet connection or through built-in Wi-Fi capability, which allows for quick, easy setup and flexibility on the shop floor. Insight CenterPoint’s software version 9.0 includes a new Library Manager tool that better organizes weldment drawings and photos; and a more visual dashboard layout with speedometer-style performance indicators. Other recent improvements include:

• Smart Part Tracking, which automatically calculates the amount of weld metal required based on an operator-input weld symbol and fillet size.
• Standard AWS weld symbols, which help save production and planning time as compared to previous versions that required symbols to be created individually by the user.

With the data from individual welding machines safety stored, management can track and assess how individual machines are performing compared to various departments or the company as a whole.Other dashboards provide graphical displays of productivity and quality. The quality dashboard tracks the number or percentage of welds made within pre-set acceptable limits for arc voltage and amperage, date-stamps every weld and identifies if quality risks have increased or decreased over a period of time. Managers can even use the collected data to gauge the effectiveness of their welder-training programs.

5.4The Fronius WeldCube Documentation and Data-Analysis System

Fronius introduced its WeldCube, a documentation and data-analysis system that connects as many as 50 welding power sources to enable accurate and continuous quality assurance and evaluation of countless parameters. Based on an industry PC with integrated software, the system is compatible with all digital Fronius machines, including the firm’s Delta-Spot resistance-welding system and with its TPS/i intelligent welding-device platform. This enables the user to document and evaluate data, including weld current and voltage, wire-feed speed, welding speed and time, arc and dynamic correction, and job numbers.

Users also can continually monitor and evaluate consumption data relating to gas, wire and energy, for example. Set values, such as job data, can be observed and recorded by the system for the entire service life of a welding system. When used in combination with Fronius’ new TPS/i welding-device platform, users can edit jobs and make comparisons across power sources.

5.5The ESAB Weld Cloud

Making its debut at the ESAB FABTECH booth was the new Weld Cloud online welding-data management system (Figure 5). It combines 3G mobile-communications technology with Wi-Fi, Bluetooth, GPS and Ethernet to allow users to alleviate firewall and connectivity issues. Running on a company’s intranet cloud, Weld Cloud is a secure, locked-down system that ensures that data remains totally confidential.

Fig.5 WeldCloud online welding data management system.

ESAB introduced its WeldCloud online welding-data management system at FABTECH, combining 3G mobile-communications technology with Wi-Fi and Ethernet to allow users to alleviate firewall and connectivity issues. The system runs on a company’s intranet cloud, and since it uses open-source software, it’s easily customized.

5.6CheckPoint from Lincoln Electric

Lincoln Electric’s cloud-based production-monitoring network (Figure 6), CheckPoint, lets users obtain performance information on their welders and welding operations located in single or multiple locations from any computer or mobile device via Wi-Fi, without the need for specialized software or IT hardware.