Real-Time Monitoring and Diagnosis of Environmental Protection Systems by Artificial Neural Networks Case study: Pharmaceutical Isolator

Abstract

: The main objective of this work is the study of risk analysis, in the field of pharmaceutical production. Some dangers can affect pharmaceutical companies’ personnel, as well as their internal and external environment, during the manufacturing process. Furthermore, the current regulations that governs this very sensitive field of manufacturing and the standards which are scrupulously very sharp. Also see the technical complexity of the industrial systems implemented. These three parameters constitute a real problem to be solved. To do this, we have developed an intelligent technique for monitoring these protection systems, in real time, in order to protect the personnel and the environment. This technique is mainly based on the use of an artificial neural network (ANN) which detects and localizes any anomalies that may occur at any time in the protection system. The experiment was carried out on an isolator belonging to BEKER Laboratories (a medecine manufacturing and development company in Dar El Beida-Algers). The test results allowed us to define the good and bad areas of the isolator operation. We concluded that, it is possible to define defaults in real time, using our new technique