System for the Diagnostics of Partial Failures in Distributed Control System Measuring Channels

Abstract

Effective control over any process is impossible without complete, reliable and timely information about the technological item condition. Such information is obtained via the automated process control system measuring channels. They include a primary transducer, a normalizing transducer, a communication link, an analog-to-digital converter, and a computational processor. Failures inevitably occur in the operation of any equipment. When a measuring channel fails, the control system obtains false information. A complete failure occurs when one of the measuring channel components fails. In case of a partial (metrological) failure, the measuring channel remains operational, but the measurement error exceeds the permissible value. The complete failure is diagnosed using the integrated diagnostics systems for instrumentation and a controller. Herewith, the integrated diagnostics systems cannot detect a partial failure, therefore, the partial failure diagnostics system should be created. It should not depend on the instrument types used. The paper considers the stages of building the crude oil distillation process regression model based on statistical data. The resulting model is the basis for the DCS partial measuring channel failures diagnostics system developed.