CONCEPTS FOR VALIDATING INTELLIGENT PROCESS MEASUREMENTS

With increased automation, operational validation is vital for effecti ve control, maintenance and management of industrial plants and proces ses. Validation should be traceable, starting from the interface to th e process up to the highest level of operations. At present, validatio n is focused at the systems level and current methods are based on con dition monitoring, and fault detection and isolation techniques using functional, analytical and hardware redundancy. With the focus at the systems level, redundancy methods exploit integration and fusion of in formation from multiple measurement devices, and thus tend to disregar d validation of the operation of the individual devices themselves. Tr ansducers and sensors provide an interface to the process and produce measurement information. Validation of the measurement output from a s ensor is imperative in situations where the cost of redundancy may be prohibitive. This paper presents concepts for validating measurements made with intelligent devices. The proposition is that an intelligent device should provide both measurement and condition information. The condition information should be used to assess the validity of the mea surement by identifying metrics which describe the sensor and process conditions, respectively. This is the lowest level of validation becau se it occurs at the interface with the process being monitored. The in formation required for this level of validation can be obtained by uti lizing all the components of the signal produced by a sensor, providin g that the sensor has a wide frequency response. An example of a valid ation approach which follows from this argument is briefly described. (C) 1997 Elsevier Science Ltd.