Mark Yseboodt, Siemens, investigates some of the cornerstone technologies of the cement industry’s digital future.
Digitalisation is a very broad concept and its applications in the cement industry are countless. At the corporate level, intelligent dashboards give a complete overview of the enterprise with just a simple mouse click. In the supply chain, extensive data exchanges between suppliers, producers and customers result in more efficient logistics. At a financial level, figures become more transparent than ever before, allowing resources to be used more efficiently, and this is only the tip of the iceberg. The ‘bread and butter’ of a cement company’s business still lies in cement production and here the use of digitalisation is ramping up rapidly. New technologies such as machine learning, artificial intelligence and digital twins are mentioned in the same breath as raw mill, kiln and cement plant.
It all starts with automation
The essence of digitalisation is to generate and analyse data to contribute to the improvement of the activities of the company by having the relevant data available in real-time wherever necessary. Zooming in on the cement production itself, a lot of process data is already available in the existing DCS and Process Historian systems. Therefore, it is very important that the automation system has an architecture that allows highly available real-time, deterministic and transparent transmission of this data, even at very high communication rates. The Industrial Ethernet standard, Profinet, can manage all data transmission requirements when applied in a correctly designed architecture. Depending on the size and requirements of the plant, Siemens has prepared several Profinet architecture blueprints to support these new designs. Security is an important aspect of the automation design and the blueprints take this into account in order to avoid any unwanted access to the plant’s network.
Nowadays, the design of such an automation network is common practice for new installations. In the cement world however, there are many existing plants that were constructed several decades ago and that now need to be modernised if they want to continue to produce for another few decades. Often the plants are uncertain about the feasibility of transforming the traditional automation architecture to a Profinet architecture, but there is no need for any concern. Siemens’ experience shows that the advantages more than justify the effort of such a conversion. For example, when a cement producer had to modernise two of its cement plants of approximately the same size, the first plant opted for the traditional Profibus technology, while the other plant decided on a modern Profinet architecture. The plant using Profinet only required 9 km of cable, whereas the plant using Profibus installed 18 km of cable – a significant difference.
With a properly designed automation network, the first step to digitalisation is completed. Plant data can now be collected, transferred and stored in a way that makes it available wherever and whenever it is required. The new architecture also supports the integration of new additional intelligence from the field to further improve the production process. Generating more data in itself is not so difficult; often it involves simply adding another sensor. However, it is important to remember that most cement plants are running plants and there is often a limit to the amount of signals that can be added to the existing control system, which was not necessarily designed to manage very large amounts of data. In fact, much of the information from the sensors is not necessarily required to control the process itself. Some details may only be relevant for maintenance purposes or for quality KPIs. In such cases, it is important not to overload the CPUs with irrelevant information. The trend today is to create an additional data channel. Smart switches route the signals through the CPU and/or through the additional channel. In this way, it is possible to add as many sensors as needed without interfering with the existing automation system. The second data channel is designed to be fully secure and the data relevant for each individual user is managed by access rights.
Paving the way for artificial intelligence
Having the data is one thing, using it to analyse and improve the efficiency of the process is another. Together with cement production experts, Siemens has developed a methodology to improve the production process, based on the analysis of real-time and historical data. This new transparent layer of data analytics is gaining importance and in combination with a high-performance DCS, the plant performance is lifted to a higher level.
Data analytics is paving the way for artificial intelligence (AI), a new technology that is finding its way into the cement industry. AI creates opportunities to solve issues that could not be solved in the past. By collecting and archiving relevant data, it becomes possible to recognise patterns and detect anomalies that previously remained invisible. With this information, plant operators can take corrective actions before an actual problem occurs. Based on the first results of this technology, cement producers can define new, improved optimisation strategies and new approaches to maintenance concepts, such as predictive and prescriptive maintenance.
Artificial Intelligence changes the view on Advanced Process Control (APC). Traditional APC solutions look at the actual measurements and try to find the optimal setpoints by changing the setpoints in very small steps. This works well for certain, well-defined processes but a kiln is a different animal. Unknown variables, such as alternative fuel types, make the traditional Model Predictive Controllers (MPC) drift over time. The result is that most of these APC systems are out of service or require intensive fine tuning. This is where AI makes a big difference because it can also consider the history of the variables and recognise patterns in these trends.
What advantage does this bring to a kiln? Based on historical data, the AI module compares the actual trends with the trends from the past and predicts what will happen in the future if the kiln continues to run with the same setpoints. What seems to be a kiln that operates flawlessly, can easily develop a faulty condition in the future and can result in unwanted alarms or even shutdowns. The AI will warn the operator in time and tell them what can happen if no corrective actions are taken. When the AI module is well-trained, it can even suggest corrective scenarios, based on past experiences. The more data that becomes available, the better the suggested scenarios will be. For example, if the plant decides to use new types of alternative fuels, the AI module will learn about this automatically. Initial tests have shown that with this solution, kiln stoppages will be significantly reduced, if not completely eliminated. The final results of these tests and the solutions linked to these will be available in the coming months. Every plant can then profit from such solutions. The financial impact will be significant.
To continue reading this article and access our May 2020 issue, click HERE: https://bit.ly/2yYrotp
Read the article online at: https://www.worldcement.com/special-reports/05052020/on-the-road-to-digitalisation/