AI in production
AI in production – How can artificial intelligence change industry?
(No Ratings Yet)
Loading...Innovative technologies are creating new opportunities where people and systems reach their limits. Artificial intelligence (AI) has long been the focus in this context. Ever since the AI-based application ChatGPT was released, the debate in this area has heated up again. Until now, it has mainly been intelligent voice control systems, such as Siri and Alexa, or self-driving cars that constituted the visible areas of AI applications. But there is also enormous potential for significant efficiency gains in industry and production. Find out here how promising the use of artificial intelligence can be in mechanical engineering, process optimisation and quality management and discover what progress has been made in weak and strong AI.
Key points at a glance
- Artificial intelligence (AI) is an area of computer science and describes the application of structures of human learning and thinking to machines.
- Along with ChatGPT, familiar areas of application of AI include intelligent assistance programmes (such as Siri and Alexa) and the development of self-driving vehicles.
- So far, AI has been used in production mainly in the fields of process optimisation and quality enhancement.
- Further opportunities to use AI are conceivable in industry and production in the future.
Definition – What is AI?
AI is the abbreviation for artificial intelligence and describes the application of structures of human learning and thinking to machines. The aim is to enable them to find answers independently without always having to reprogramme them. AI is therefore also referred to as “machine learning”.
According to Wikipedia, https://en.wikipedia.org/wiki/Artificial_intelligenceAI is an area of computer studies that is subdivided into strong and weak AI.
What is strong AI?
Strong AI refers to computer systems having the capacity to work in the same way as human beings to complete difficult tasks. Despite decades of research, strong AI is still a vision that has not yet been achieved completely.
The current view is that strong AI will have a different type of cognitive structure from human beings and will not be comparable with the evolutionary stages of human thinking. At the same time, strong AI will probably not have feelings, although it could be able to simulate them.
What is weak AI?
Weak AI is about overcoming specific application problems. The intention is therefore to support human thinking and technical applications, not to replace them. Ultimately, it is thus not a question of creating consciousness but of simulating intelligent behaviour, which can be achieved using methods of mathematics and informatics through so-called machine learning.
The ability to deal with uncertainties and to learn are among the important factors that an AI system should have. In contrast to strong AI, significant progress has been made in the area of weak AI in recent years. Leading examples include voice recognition, the developments towards self-driving cars and the Google algorithm. It is therefore weak AI that is currently already being used in production and industry and will grow in importance in the future.
ChatGPT as a modern example of AI
The launch of ChatGPT by OpenAI has triggered plenty of debate and turned the focus back on AI (artificial intelligence). The future development of the software is not entirely predictable. What seems certain, however, is that it will have an effect in the areas of education and communication. Further programmes are required in production and industry, however. We would like to take a brief look at this area now and report on the latest developments.
What is the significance of AI in industry?
AI in industry is data-driven. Already a large volume of data is being generated, processed and analysed. In a production environment, these volumes of data can form the basis for generating digital copies of entire production systems. These digital twins then help to make the planning and design of machines and the entire production process faster, more flexible and more efficient. The data of the digital twins and from the IoT (Internet of Things) is analysed closely in this context and used for optimisation, simulation and decision-making. IoT applications use machine learning algorithms to analyse the huge data volumes.
In this way, systems can be continually adapted to new circumstances and further optimised without human intervention. With increasing networking and continuous analysis, the AI software can learn to identify complex interrelationships in systems which a human being cannot see.
From the current perspective, this is all still very complex, as simply processing large data volumes is a big challenge in itself. But this is precisely the prerequisite for achieving “learning progress”.
What examples of AI are there in industry?
Intelligent software is already available and digital companies in industry are now a reality. Artificial intelligence is used in a variety of ways, such as in language recognition, for processing simple orders, as a virtual assistant and in warehousing and logistics. About 62% of large companies were already using AI technology in 2018, while for small and medium-sized enterprises there was – and still is – a need to catch up.
How is AI used in production?
A future in which transporters drive through factory halls independently, systems optimise themselves in relation to energy consumption during operation and production machines carry out quality controls even during the production process and make direct adjustments as necessary is not a vision but already reality in some cases.
An expert opinion based on practical experience
“My personal opinion is that real AI begins with self-learning machines. But there is a multitude of programmable applications that are already achievable today and can be used in mechanical engineering. A transporter that drives independently through factory halls may already be an example of artificial intelligence (if the transporter finds its way itself and can work out new routes independently) or it can be a machine that is equipped with sensors, distance measuring devices, etc. and only covers certain routes. There are good reasons to use either variant. It is our task as machine designers to find the optimal solution for our customers.”
(Hartmut Geyssel from Geyssel Sondermaschinenbau & Konstruktionsbüro in Cologne)
Production must become more efficient, faster and cheaper. This can already be put in place today. But in mechanical engineering and production, the reservations about the use of AI are still relatively big. First, parallel systems and digital twins must usually be set up to analyse how they work with the help of AI systems. This is because it is not usually possible to learn from mistakes in real production practice. Second systems must therefore be used for “learning” and verified over an extended period with the aid of human intelligence. This process is complex and takes time. This is precisely why many machine designers and production companies have so far found it difficult to assess the actual benefits. The opportunities are therefore present, but the feeling is that success is some way off.
What examples of AI are there in production?
Digitalisation has now become an integral part of production and AI is one of the most important technologies in relation to digitalisation and therefore a big opportunity for manufacturing companies.
In production, AI could lead to long-term cost benefits. If, for example, more production has to be relocated to Germany in the future (because the coronavirus crisis made everyone realise how painfully fragile the supply chain is), cost efficiency will be very important. So it is worthwhile taking a look at some real examples of applications from practice.
The following examples of areas of application for AI in production illustrate the wide range of opportunities for using AI in this context:
Use of robots in production
Germany is the most automated economy in the European Union, with around 230,000 industrial robots. The International Federation of Robotics reported this when it presented its latest annual report “World Robotics 2021”. The number of newly installed robots in Germany reached 22,300 units in 2020. That is the third-highest figure ever achieved, despite the fact that it was a crisis year. The new robots now largely work hand-in-hand with their human colleagues. So far, the robots are not particularly intelligent. That could change with the increasing use of AI, as in theory robots are extremely capable of learning and are able to take over more and more cognitive tasks.
Maintenance with AI
Through so-called preventive maintenance (as part of Total Productive Maintenance), it is possible to determine the optimal maintenance time based on wear on the operating equipment using algorithms, to plan maintenance intervals better and largely to avoid production downtimes. Although a great deal of automation is already taking place in the area of maintenance, strong AI is not yet being used. That could change in the future.
Quality control with AI
AI-supported methods are becoming increasingly important in quality management, especially “computer vision” and “machine learning”. These help with early identification of product damage of a sort that human beings often cannot spot in manual checks.
Offsetting the shortage of qualified personnel
In an age when there is a shortage of qualified personnel in many areas, intelligent assistance systems can help. Semantic AI technologies are used most widely in this context, which help employees to take decisions more quickly and efficiently and complete tasks in greater quantity and better quality on the basis of the available data.
In addition, intelligent machines can identify discrepancies in the production process at an early stage by analysing vibrations and noises and draw attention to them by issuing warning signals. This saves on waste and production downtimes.
Other important questions about artificial intelligence in production
Is AI worthwhile for medium-sized businesses?
Process control by self-learning systems is currently very limited in the production environment, at least in small and medium-sized enterprises (SMEs). Although 25% of large companies and 15% of SMEs essentially already use AI-based applications, the figure for small and medium-sized enterprises in the area of production falls to just 8%. There is still a lot of unexploited potential here. This is particularly concerning, but growth in German GDP is forecast to be 1.2 percentage points a year up to 2030 as a result of the use of artificial intelligence alone. The manufacturing industry therefore has opportunities to develop through AI, including medium-sized businesses.
What are AI-supported processes?
Production can be optimised by AI-supported processes. Various AI functions and highly developed analytical methods are used to optimise productivity, quality, safety and output, while simultaneously minimising downtimes. Through self-learning systems, errors in the production process and defective products can be identified automatically and more quickly and can thus be avoided.
There are already intelligent machines of this sort, but it is necessary to assess in each case the extent to which their use in the specific company is expedient and efficient.
What does robotics mean in the context of AI?
Robotics, also referred to as robot technology, focuses essentially on the design, control, production and operation of robots that can take on tasks in production and operation. Findings from mechanical engineering, electrical engineering, computer science and – to a growing extent – from artificial intelligence (AI) feed into robotics and further improve the capabilities of robots.
What does Industry 4.0 mean?
Industry 4.0 is the term used for a project that is intended to take the future of production into a new age through comprehensive digitalisation. Industry 4.0 is therefore the start of the fourth industrial revolution. The first industrial revolution was shaped by steam power, the second by conveyor belts (which enabled mass production) and the third by the use of electronics and IT (e.g. CNC machines). The fourth revolution is about to extend the series and optimise industrial production through networking, transparency and assistance systems.
Summary
The use of AI in production is becoming increasingly important. Alongside the big industrial companies, small and medium-sized enterprises will be able to – and will have to – exploit that potential in the future. In the long term, it will be possible to organise production processes with AI so that they are significantly more efficient and more cost-effective. The quality of products can be improved in this way, while the burden on skilled personnel can be eased and they can be supported.
We at GEYSSEL are always at the cutting edge of development thanks to continuous internal training and collaboration with research bodies and universities. We also work closely with software developers. We will be happy to support your company on its journey into a future with artificial intelligence.
FAQ
What is strong AI?
Strong AI refers to computer systems having the capacity to work independently in the same way as human beings to complete difficult tasks. Strong AI is currently still a vision and has not been fully realised
What is weak AI?
Weak AI deals with specific application problems. Human thinking and technical applications are simply supported by AI, not completely replaced. Ultimately, it is not a question of creating consciousness with weak AI but of simulating intelligent behaviour.
Where is AI used in production?
AI is used in production above all in the areas of quality control, maintenance and specific applications for robots. Skilled human personnel are supported and the burden on them is eased in this way.
What does Industry 4.0 mean?
Industry 4.0 is the term used for a project that is intended to take the future of production into a new age through comprehensive digitalisation. Industry 4.0 is therefore the start of the so-called fourth industrial revolution.
