Vaccine manufacturing is subject to strict regulatory requirements to ensure that vaccines or other manufactured medicines are as safe as possible. In order to be able to produce vaccines at an adequate quality, Novavax needs to have a stable operation of its production technologies. Since the medicines are manufactured in classified clean rooms according to European and national regulations and ISO standards, it is also necessary to have a stable operation in this respect, which is ensured by the air handling system.
In the event of an unforeseen malfunction, such as a jammed fan bearing or drive motor during a production campaign, the environmental parameters may deteriorate. All production batches in progress must then be subsequently disposed of. Failure to comply with clean environment parameters thus represents a significant financial risk. Therefore, in addition to standard preventive maintenance, the company decided to implement predictive maintenance technology.
Neuron Soundware has equipped 4 air handling units with predictive maintenance technology in a set of an electric motor for the drive and impeller of the air handling unit on a common bed, enclosed in a soundproof cabinet within the air handling plant. The solution ensures the collection of sound data by sensors that are connected to an IoT device (nEdge), where this data is then processed.
A complex machine learning algorithm in nEdge analyzes the collected data. nEdge is connected to the internet and the data is displayed through the nGuard cloud portal. The project started by recording the nominal sound and vibration of the monitored devices and training a mathematical model for a comprehensive anomaly and vibration detection service.
Now in case of any anomalous sound (compared to the nominal machine behavior) the system will alert the operator. In the case of normal behavior, the online reporting in the portal shows the standard operation and confirms that the machine is fine.
If a deteriorating trend is observed on the equipment, maintenance capacity and spare parts can be provided in advance. Subsequently, a shutdown can be planned with the specific production department according to the production plan and the problem can be solved.