October 2006

In this issue

	Classes
	Trick of the month: Is this cycling normal?
	To contact us

Coming classes and seminars

Detailed Loop Analysis with ExperTune
October 16-18, Hartland, WI

ExperTune Advanced Tools
October 18-20, Hartland, WI

Seminar on modern control methods
October 31-November 2, Vancouver, BC and
November 29-December 1, Bangkok, Thailand and
December 5-7, Malaysia

Control Valve Selection
November 14-16, Moncton, NB

PID Loop Tuning
November 21, Ottawa, ON

Introduction to Instrumentation and Process Control, Process Control Basics for Non-Specialists
November 22-23, Ottawa, ON

See our calendar for 2006

This month's tip : Is this cycling normal?

You observe on the screen that a temperature (PV) is cycling. Is this normal or is there something wrong? Is it noise, disturbance or something else?

In process control, cycling is abnormal and should be eliminated since oscillations have negative effects. (They increase energy costs, disturb other parts of the process, reduce quality, reduce equipment life, etc.)

Noise or disturbance

In a control loop, the controller uses the error between SP (set point) and PV (process variable) to manipulate CO (controller output) to bring back PV to SP. The settling time for a control loop is the time needed to bring back PV to SP (usually we consider the time it takes to reach 95% of the change). This settling time depends on the process and the controller tuning parameters.

If the loop is tuned sluggishly, the settling time is long and the error needed to bring back PV to SP is large. If the loop is tuned aggressively, the settling time is smaller and the error needed to bring back PV to SP is smaller, but temporary cycling could occur (damped oscillations).

If oscillations (or repetitive phenomenon) occur at a period that is shorter than the settling time, the controller is not able to fight them; in that case, we call them noise. Otherwise, they are disturbances.

To reduce or remove disturbances, we have to find their origin and eliminate them, or at least change the rate. For example, if a solenoid valve is used to reverse the flow in a heat exchanger that reduces fouling, the rate should be adjusted to reduce its impact on other loops. If a reservoir level is controlled by using hi and low limits and this cycling induces oscillations in the process, one should consider replacing the On-Off scheme by a controller and a modulating valve (or variable speed drive).

To reduce or remove noise, we also have to pinpoint its origin. If the noise cannot be eliminated, we must add a filter on the PV in order to reduce the impact of this noise on the controller output and its propagation to other loops.

Loops without oscillations

In processes in which all loops are optimized, all PVs should present normal statistical distributions, have a flat power spectral density, and an autocorrelation plot remaining within confidence limits after settling time. There should be no cycling or white noise (all frequencies are equally distributed).

Rules of thumb:

ts >> td	or	ts ~ 10td
<< td	or	~ td ÷ 5

ts	settling time	time to reach 95% of the total change
td	dead time	time to observe a change on measurement       after a controller output change
	first order filter time constant

For example, if a temperature loop has a dead time of 1 minute, the filter time constant should be around 12 seconds and the settling time should be around 10 minutes. An oscillation that has a period of 1 or 2 minutes will be noise, while an oscillation that has a period of 30 minutes will be a disturbance.

To contact us

	Lévis's office	+1 (418)834-4321
	Pointe-Claire's office	+1 (514)695-3492
	Hartland's office	(877)867-6473
	Toll free	(877)867-6473
	Email	info@topcontrol.com