Helping the process industries find a way to Industry 4.0

Digital plant monitoring for predictive maintenance based on digital twins. ©Fraunhofer IFF.

As far as the process industries are concerned, the road to Industry 4.0 is still long. Fraunhofer researchers and engineers are working on interconnecting process systems so that they can be serviced and maintained predictively, by combining operating data with employees’ knowledge.

A lot of time is lost compiling relevant information and documents, or gathering experienced employees’ knowledge, during the troubleshooting of process systems. Important know-how from maintenance and manufacturing staff is also tremendously insecure because it is unavailable whenever employees are ill, or is lost to a company entirely when employees depart. It would therefore be desirable to have it constantly available for automated system control. This is where Industry 4.0 solutions can help.

Industry 4.0 is still in its infancy in the process industries, though. There are only sporadic research projects. This is why many companies in the chemical, pharmaceutical, steel and cement industries, as well as their suppliers, fear falling somewhat behind in technological development. Researchers and engineers are developing a new digital monitoring system in a project at the Fraunhofer Institute for Factory Operation and Automation IFF, which will enable the process industry to use Industry 4.0 technologies. It is intended to simplify process system maintenance and servicing significantly in the future. The researchers and engineers are doing this by digitising system monitoring and interconnecting every relevant level of a process operation in several ways.

The researchers and engineers are using a fluidised bed granulation plant as their technology demonstrator. Such plants produce products such as granular pesticide. “The envisioned interconnectivity of systems is based on their digital twin,” explained Dr Nico Zobel, a research manager at the Fraunhofer IFF.

The process being developed by researchers will interconnect plants in three dimensions for maintenance. The first dimension spans the lifecycle. The experts use plant engineering documents (eg, the three-dimensional CAD model created as the plant was engineered) for equipment operation. Workers needing information on a particular component, such as a pump, can scan the pump’s QR code using a tablet computer on which every available planning document on the component is displayed. Workers can additionally view the pump’s stored operating data, such as temperature and pressure curves. The digital twin also helps with troubleshooting: an interactive recommended action can be issued for every problem the control system reports. This will guide employees as they localise problems, and digital instructions will tell them how to eliminate problems step by step.

The second level of interconnectivity the researchers and engineers intend to implement is vertical interconnectivity. “The sensors installed in a plant send the data they collect to the cloud. Any data can already be incorporated in the planning of maintenance actions at this early stage,” explained Dr Zobel. This makes it possible to implement predictive maintenance in such process plants. Injectors such as those found in granulation plants are an example: injectors clog from time to time, thus increasing the likelihood of bringing a plant to a standstill. The more sensors that send their data to the cloud, the more precise the base of data is, which the system then uses to ascertain the next scheduled injector servicing.

The researchers and engineers not only take operating data as the basis for the second level of interconnectivity, they also combine it with employee know-how. The researchers and engineers ask the employees specific questions in order to collect their know-how. They use the responses to their surveys to develop a mathematical model of probabilities of wear or failure. They additionally link this model with artificial neural networks, which are used to develop correlations between sensor data and a component’s wear allowance based on the system’s historical data. This provides a basis for delivering good forecasts of the future performance of individual plant components.

The third level of interconnectivity is intended to link current production with the supply chain. If, for instance, a seal in equipment has to be replaced, employees are instantly notified if it is in stock. If it is not, the purchasing process is started automatically.