In today’s world an enormous amount of data is generated every second which, when combined with advances in affordable data storage and increasing computing power, provides a great potential to take advantage of. Completely automated reporting and the use of artificial intelligence (AI) will be shown as examples of how to capitalise upon this potential in a production location that has demonstrated its value for over 100 years. This shows that these techniques are not limited to plants that are brand new.
Artificial intelligence is a domain in which intelligent programs can solve various problems. Machine learning and deep learning are part of AI, where powerful algorithms can be developed that can learn from available data and improve process control. Further, when the algorithm is exposed to new data it also learns from this, creating a self-reinforcing beneficial loop. These algorithms can be used, among others, to automatically detect if employees wear their safety gear; to predict in-line product quality parameters or to predict machine health in the future. Many more possibilities exist.
As an example, an algorithm will be presented that can predict a product quality parameter. This algorithm can predict the moisture content in granular fertilisers; one of the key parameters in granular fertiliser production, with a significant impact on the gas consumption of the production process. The output from this algorithm can be connected to the plant controls to automate this process. Alternatively, this information can be shown to the operators to help them improve their own decision making.