It’s become increasingly clear that in order for producers to compete more effectively in the marketplace, they must become more efficient. This means making better decisions and being smarter with what they do.
For most manufacturers, labour is still the single biggest cost of production. So, any technique or system that can minimise labour content will generate obvious cost benefits.
The logical way of reducing labour is to mechanise the production process so that machines effectively do the work. Machines can work 24/7, without getting tired or needing a break or holidays. Also, once properly set up, they can produce high-quality output consistently.
This process of mechanisation has been underway for many decades and continues unabated. It has led to a marked increase in efficiency, better output quality and reduced production costs. It has also allowed products to become more intricate, as customers demand ever greater diversity in goods.
Since around the 1970s, mechanisation has incorporated computerisation. And as computers have become faster and more powerful, this has in turn allowed machines to become increasingly ‘smarter’, as they take on a growing number of decisions, for tasks that are ever more complex.
However, the computers used throughout industry (i.e. industrial controllers such as PLCs) still make their decisions based on a very rigid, pre-defined program that executes cyclically. The program is designed to handle a fixed set of input conditions, with little flexibility and without any ability to adapt or change what it does.
While such control systems are effective and widely used, their scope for decision making is relatively restricted, limiting what they can do. They also have no way of handling unforeseen or random events, meaning some human intervention is still required.
But perhaps the biggest limitation of traditional controllers is that any change in functionality must be implemented manually by a programmer. If producers are to remain competitive, it requires constant improvements in processes, and these controllers offer little to no help in this.