Chemical processes usually impose severe constraints in the fields of safety, reliability, efficiency. The know-how, necessary for their control and supervision implies numerous and wide domains of knowledge (chemistry, data treatment, process control, signal processing, etc.) of both quantitative and qualitative natures. An integrated implementation of the latter requires the development of dedicated tools and techniques for real-time supervision of chemical processes. The paper presents a few aspects of an expert system for real-time supervision and fault detection applied to the industrial wastewater detoxication plants (WWDP) for surface treatment industries. The goal of the project is the design of information and decision support systems, in order to improve the operation (safety, reliability, efficiency) of WWDPs as well as to provide an efficient training tool for the personnel (staff, future operators, etc.).