Cycling of Closed Loop Control Systems

Closed-loop control systems can waste energy and shorten equipment life when they switch on and off too frequently around a set point.

Use this page for

Heating systems controlled by thermostats.
Pressure-controlled compressor systems.
Lighting systems controlled by ambient light level.
Any process where frequent short on-cycles suggest poor control behaviour.

Recommended loggers

Logger	Best fit	Measures
EC-7VAR-RS Electrocorder EC-7VAR-RS Three Phase Voltage, Current & Power Factor Recorder	Use when cycling may be linked to both supply voltage and load behaviour.	Three-phase voltage, current, power factor and energy.
EC-3V Electrocorder EC-3V Three Phase Voltage Recorder	Best where three-phase supply behaviour may be contributing to cycling or resets.	Three-phase voltage.
EC-1V Electrocorder EC-1V (Mailable) Voltage Recorder	Best for compact single-phase voltage logging where the control signal or supply voltage needs to be recorded.	Single-phase voltage profile.
LS-1V Electrocorder LS-1V Interruption Voltage Logger	Best for cycle counting, voltage interruption and on/off timing proof.	Single-phase voltage, interruptions and duty-cycle timing.

What closed-loop cycling is

Frequency Distribution of On/Off Events in a Closed Loop Control System - Voltage Logging

On/off cycling can be a real problem in closed-loop control systems. It occurs when a control system repeatedly switches equipment on and off around a set point, rather than making fewer, longer, more useful control actions.

Common examples include:

Thermostatically controlled heating systems.
Pressure-controlled compressor systems.
Ambient luminance controlled lighting systems.

Measured heating example

Electrocorder monitored the thermostatically controlled oil-fired central heating system in the Acksen office building over a period of about six months. The results showed a large number of very short boiler-on events, all requested by the thermostat controller.

The boiler came on 1,363 times for two minutes or less, and 255 times for between two and four minutes. When the cumulative boiler-on time was analysed, 33% of the boiler run time came from on-periods of 10 minutes or less, and 40% came from on-periods of 15 minutes or less.

For short on-cycles of 10 minutes or less, a boiler cannot add much useful heat to the building. Much of the fuel is used to start and warm the boiler, then lost in the boiler and surrounding area. From the recorded data, eliminating boiler demands of 15 minutes or less was estimated to offer a saving of around 30% to 40%.

How the cycling was reduced

The practical solution was to add a delay-on timer. The timer receives the thermostat on-signal, waits for 10 minutes, and only then passes the signal to the boiler. This removes short thermostat oscillations around the set point while still allowing genuine longer heating demands to turn the boiler on.

Flowchart for monitoring cycling levels in a closed loop control system using an LS-1V Electrical Data Logger

Two LS-1V data loggers were used to monitor the revised system. One logger recorded the demanded boiler-on signal, and the other recorded the actual boiler-on signal after the delay timer.

Boiler run-time results:

Cumulative demanded boiler-on time: 21:02 hours.
Cumulative actual boiler-on time: 9:57 hours.
Demanded on-time not given: 11:05 hours, a 52% energy reduction over the trial period.

Boiler cycle-count results:

Number of demanded on-events: 184.
Actual number of on-events: 25.
Demanded on-events not given: 159, an 86% reduction in on-cycles.

Projected annual cycle-count results:

Average number of demanded on-cycles per year: 8,375.
Average number of actual on-cycles per year: 1,156.
Demanded on-cycles not given: 7,219, an 86% reduction in on-cycles.

The measured trial achieved a 52% reduction in fuel consumption over the test period.

These figures are not theoretical or speculative. They came from a real working heating system, with both the thermostat demand signal and the actual boiler-on signal recorded.

Comfort and control considerations

A delay timer does introduce a delay between the demand for heat and the supply of heat. If the demand is a real temperature demand, the internal temperature can continue to fall slightly during the delay. When the boiler then starts, it may run for longer to bring the building back to the set point.

Even with this effect, the measured office trial still achieved a 52% saving. There was no noticeable effect on internal office temperature during the timer trial, but this should always be checked on the real system being controlled.

The timer effectively introduces a dead band. In the recorded trial, 56% of the demands for energy were not genuine temperature demands, but thermostat oscillation around the 21C set point. Reducing run time by 52% and on-cycles by 86% also has a significant effect on boiler life.

Compressors, pumps and lighting

Compressor cycling can also be a problem in pneumatic compressor systems, where pressure control attempts to maintain a single set point. A better method is often a dead-band system: a minimum pressure starts the compressor, and a higher maximum pressure stops it.

If a delay timer is used on a compressor system, the delay must not be so long that air demand depletes the receiver before the compressor is allowed to start. The minimum working pressure of the equipment fed by the reservoir and the maximum air demand both need to be considered.

Pumps and compressors are also affected mechanically by repeated starts and stops. Reducing the number of on/off cycles can improve system life as well as energy efficiency.

Controlled lighting can cycle when the set point is a lux level. Some systems use separate on and off lux thresholds to create a dead band. Lighting cycling normally happens around dawn or dusk, so the potential improvement is usually more limited than in heating or compressor systems.

Using Electrosoft for cycle analysis

The LS-1V was designed for cycle analysis, but Electrosoft can perform cycle analysis with other Electrocorder logger data too. For voltage loggers, define an on-voltage and off-voltage threshold. For current data, define an on-current and off-current threshold. Electrosoft will then calculate the cycle behaviour from the recorded data.

Related guides and products

Machine Trip and Reset Loggers: for equipment trips, resets, lockouts and intermittent electrical faults.
Voltage Problem Loggers: for high voltage, low voltage, voltage variation, dips and interruptions.
Load Survey and Current Loggers: for feeder checks, current logging, load surveys and capacity planning.
Energy Audit Loggers: for energy audits, load profiles, peak demand and before/after savings proof.
All application guides: choose an Electrocorder by the job or problem you need to investigate.

Relevant products:

EC-7VAR-RS: Electrocorder EC-7VAR-RS Three Phase Voltage, Current & Power Factor Recorder.
EC-3V: Electrocorder EC-3V Three Phase Voltage Recorder.
EC-1V: Electrocorder EC-1V (Mailable) Voltage Recorder.
LS-1V: Electrocorder LS-1V Interruption Voltage Logger.