Browse Prior Art Database

TECHNIQUE TO DETERMINE OPTIMUM PERIODIC SAMPLING

IP.com Disclosure Number: IPCOM000241743D
Publication Date: 2015-May-27
Document File: 3 page(s) / 12K

Publishing Venue

The IP.com Prior Art Database

Abstract

The present invention proposes a technique to determine optimum periodic diagnostic sampling of many valves in a plant process. The technique includes software that arranges every valve in the plant with required tag-names and addresses necessary to connect to at least a set-point, position and pressure, in a list. Available number (N) of connections that are made to field devices per hour is then determined. According to an implementation of the technique, a portion of the connections is assigned to the list. According to one embodiment, the technique includes a sub-list that samples at intervals (Is), for example, of two minutes. The number of valves in the first sub-list is computed as N/Is.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 3

273248

TECHNIQUE TO DETERMINE OPTIMUM PERIODIC SAMPLING

BACKGROUND

The present invention relates generally to control valves in a process plant, and more particularly to a technique for determining optimum periodic sampling in the process plant.

Diagnostic data from control valves in a process plant is used to guide predictive maintenance of the valves. The diagnostic data may be obtained by connection to a valve positioner via known communication protocols running through an interface with a control system.

In a process plant there may be numerous valves, in the order of hundreds and thousands, which require monitoring. The communication protocols can deliver diagnostic data samples at a rate ranging from 1 second to 5 seconds per sample. Some diagnostics require high frequency sampling, but others may not require sampling as frequently and relatively longer sampling intervals are sufficient. For example, friction is a particularly important diagnostic measure of a valve's condition. Measurement of friction is best accomplished by observing a set-point, position, and actuator pressure from the valve positioner while the valve moves in one direction, for example, opening a small amount, then moves in the opposite direction, and then reverses direction of movement at least one more time, i.e. the valve undergoes periodic motion.

Observing movement of valves at high frequencies consumes communication bandwidth. An optimum use of bandwidth can be accomplished by sampling for diagnostic data during cyclic periods at which the movement of valve(s) is known to occur. During such periodic movement, recording a few diagnostic data samples during each movement is typically sufficient for obtaining frequency characteristics of the movement of various valves. It is known that control loops often provide a damped frequency of oscillation in response to disturbances, for example, to minimize the disturbances, at the frequency characteristics or the cyclic periods at which the movement of valves occur. Usually, a large number of valves are observed using relatively slow communication protocols. Also, some valves move very slowly. In order to measure friction in such valves, at least two reversals of movements are monitored. It would be sufficient to monitor the valve about 16 times or less, during period of characteristic oscillation.

A conventional technique includes an online valve diagnostic (OVD). The OVD connects to a single valve at a specified time or interval and acquires a predetermined number of samples at a rate permitted by a data carrier detect (DCD) input/output (I/O), which is once in every 1 to 5 seconds. However, advantage of monitoring control valves is limited to valves with fairly rapid cycles. Frequently certain valves yield


Page 02 of 3

273248

insufficient information because they are not observed over a long enough period to capture data from several rever...