Energy saving by voltage reduction

I have no idea what this should have to do with single- or 3-phase operation mode. To my knowledge such techniques are only applied to single-phase loads, but of course the voltage reduction devices are made up in 3-phase design because they can be economic only where greater numbers of these (smallish) loads are being operated. Then, as a rule, they are (more or less) evenly distributed across the 3 phases.

But it remains to be considered that different loads react in totally different manners to the application of this technique:

• Incandescent lamps: Yes, there is a savings potential, although the drop of power is somewhat less than by the square of the voltage, as you would expect from an ohmic load, because the resistance varies with temperature. In a cold state an incandescent lamp has only ≈10% of the resistance it has at normal operation. But: The efficiency drops drastically with voltage (Fig. 1)! Although the lifetime expectancy increases by about the same factor, the application is not recommended because under normal conditions the efficiency of an incandescent lamp is already very poor.
• Other ohmic loads: What could these be? Only heaters, such as cookers, ovens and so on. These are usually thermostatically controlled, so a reduction in input power by reducing the voltage will result in a shift of the on/off ratio of the thermostatic switch if it is of the plain and simple bi-metallic type. Electronic control will have the same result. The heat-up time becomes longer, and the average power intake across the whole heat-up and subsequent keep-warm period remains the same, i. e. only disadvantages.
• Fluorescent lamps with magnetic ballasts: This is the only application where the voltage reduction is very efficient and highly recommendable. The losses in the ballast drop by the square of the current, while the lamp efficiency slightly increases with reduced current. At the same time the voltage drop across the lamp increases when the current decreases, so the lamp power drops under-proportionally while the ballast power (loss) decreases by the square of the current. This yields a triple effect in favour of better overall luminaire efficiency.
• Fluorescent lamps with electronic ballasts: Here there is usually no visual or palpable effect because electronic ballasts are commonly made up in a way as to offset any input voltage variances. This is in principle an advantage but excludes them from any opportunity to influence their operating behaviour by varying the line voltage. Rather, since the output power is kept stable, the input current will increase by the same factor as the voltage is reduced (Fig. 2). Load on the mains becomes heavier, losses rise, and the number of ballasts that can be fed from one circuit drops.
• Compact fluorescent lamps: These are basically small fluorescent lamps with integrated electronic ballasts, but the ballasts are of a plainer and cheaper design. Mostly their electronic control is merely designed to come from a square to a linear voltage dependence of the input power, not right through to a stable input power.
• Other electronic equipment (PCs etc): You could as well group electronic ballasts in here. Like them, a PC or any other electronic circuit requires a certain power to fulfil its function, and in order to supply the required power to the electronic circuitry the usually applied switch-mode power supplies will draw a higher current from the mains when the voltage is lower. In fact there are already many switch-mode power supplies around which are designed for universal use at all line voltages around the world (from 90 V to 264 V, including the permitted tolerance levels of ±10%). The current ratings usually reflect that the power has to remain constant (Fig. 3). No point in voltage reduction, only disadvantages, apart from a slight mitigation of relative harmonic currents with lower voltage. But since TRMS current is higher, the absolute harmonic currents tend to be higher rather than lower.
• Induction motors: Don’t do it! As with electronic devices, the power demand at the shaft is practically the same at lower or higher voltage, since the rotational speed is mainly determined by the frequency, hardly by the voltage. So current will increase. Remember that the starting torque is proportional to the square of the voltage. Normally the losses in an induction motor will increase with both over- or undervoltage. Now the motor and the driven system must be designed in a way so that, when the voltage is within the tolerance levels, the system works properly. Yet, when you decide to reduce the line voltage only within the tolerance range, it remains to be considered that the start-up current of an induction motor is about 7 times the rated current and that the voltage reduction gear introduces additional impedance into the circuit! So the voltage may very well drop far below the lower tolerance limit at the moment when the motor is switched on, the motor may stall, the (potentially vital) driven system will fail to start, and the current will persist in the start-up current range! Therefore not just no point in reduction, but even dangerous!

Summary:

The voltage reduction technique is a very good means of making fluorescent lamps with magnetic ballasts more efficient. In fact at 207 V a 58 W lamp with a class B1 ballast is already more efficient than with an electronic class A3 ballast! Similar effects may eventually be expected in street lighting (high pressure discharge lamps).

With incandescent lamps a certain energy saving is possible if a substantial loss of light output is accepted. Lamp life can be substantially extended.

With nearly all other applications the effect is either futile or adverse, may even become dangerous! So the voltage reduction technique should only be applied where dedicated lines for the lighting are provided and the type of ballasts used is known.

Regards, Stefan

Great comments from Stefan&Nathan

Nathan wrote "but you forgot to mention that many people who fit voltage reduction equipment do so for an entirely different reason"

But Nathan himself forget or somehow scare to say for following further reduction ( of course safely) in voltage reduction. You can find many documents for example for 1% voltage reduction from the nominal, i.e one can check it from:

Statistical Test of Energy Saving Due to Voltage Reduction Kirshner, D.; Giorsetto, P.
IEEE Transactions on Power Apparatus and Systems Volume PAS-103, Issue 6, June 1984 Page(s):1205 - 1210
Digital Object Identifier 10.1109/TPAS.1984.318450

Summary: A number of utilities have conducted field tests of the effect of distribution voltage reductions on energy consumption. The tests have been interpreted as giving mixed results. This paper reviews eight tests and statistically analyzes the results. These tests consistently support significant energy savings as a result of voltage reduction. Percent energy savings for each 1% reduction in voltage average 0.76%, 0.99%, and 0.41% for residential, commercial, and industrial class loads, respectively.

Fan Loads & Power Regulations For High Profits

In cities like Delhi where I live peak demand varies between 2,200 MW to 4,200 MW depending on the season and even during peak summer season rains due to local depression could bring down the cooling load and avoid irrigation load – mainly fans, air conditioning and tube wells, power demand can be managed well by reducing voltage in the distribution network.

Fans load which is significant responds very well reducing power consumption by lowering supply voltage. Theoretically 10% reduction in rpm results in 30% lowering of load when air circulation is reduced by 10% only. Whenever there is power cut and fans are operating on inverter I turn the regulator one step lower that is enough to bring down load on inverter by 25% though reducing air delivery by about 10%. Thus reducing load on inverter and extending its battery life.

Utilities in Delhi too like UK tends to supply power at 240V-250 voltage range against 230V specified to pump more units in to consumer bills in affluent areas where every additional unit billed is charged at highest tariff rate and at 200V in low income areas thereby reducing units billed where tariff rates for saved units are substantially lower. (There are three tariff slabs in Delhi and most of India for domestic power.)

Utilities use voltage regulations effectively to bill more units at highest tariff and simultaneously reducing sales in low tariff zones.

Ravinder Singh, Inventor

Ah yes, another version of the perpetual motion machine. As described in the original response, energy reduction (and corresponding reduction in output) by voltage reduction is simple and easy, the big question is whether efficiency improvements can be achieved by voltage reduction.

The primary case described is that of the magnetic fluorescent ballast. Light output and energy use are both decreased by voltage reduction. Lower light output is often not noticeable but of course you don't need a fancy voltage reduction gadget to achieve this and at some point it most certainly is noticeable. What is more interesting is the claim that energy efficiency of the fluorescent magnetic ballast systems is improved by the voltage reduction. This claim is made on the basis of improved efficiency of the fluorescent tube operating at lower voltage and of reduced ballast losses at lower voltage.

Fluorescent tubes do not necessarily improve efficiency when operated at lower voltages. They operate most efficiently at a particular power level, corresponding to a particular coltage and current, and at lower efficiencies at both higher and lower power levels. In the particular case study (funded of course by a magnetic ballast manufacturer lobby group), the common 58W T8 tube was used and shown to operate slightly more efficiently at below 58W. Most tubes do not show this relationship, in fact many of them operate more efficiently at significantly increased power levels. When either lower light output is desired from an existing number of fittings, or when improved efficiency can be achieved by lowering power levels, ballasts with a ballast factor below 1.00 are available.

The second factor is the reduced (magnetic) ballast losses at lower input voltage. These are described here as proportional to the square of the current which is a little simplistic, nevertheless ballast losses decrease when the ballast is operated at a lower power. A different way of achieving lower ballast losses would be to use a more robust ballast (thicker windings, etc.) to achieve lower losses at the same power. The case study I have seen (again funded by the copper sellers) plays a little fast and loose with such things as the power factor correction and the measurement of the real and apparent ballast losses. Be very careful since in most application you will need power factor correction and will not achieve the savings shown in the study.

In summary, energy efficiency savings by this method can be made when applied to certain types of existing equipment, not to others. In all cases, the best energy efficiency will be achieved by appropriate choice of equipment to operate at your supplied voltage, not by adding additional equipment in front of sub-optimal appliances. Whether and when an optimal solution should be chosen will depend on retro-fit costs vs the payback period.

We are not promoting any perpetual motion machine. We do not promote any specific machine at all. Where did you read we do?

Indeed the big question is whether efficiency improvements can be achieved by voltage reduction. However, we did not want to ruminate this question but rather provide you with an answer, and this can only be done by a qualified measurement. Believe us, we have been searching for such measurement results virtually for years now, but either this has never ever really been measured, or the results were such that the corresponding organisations were displeased to disclose the results. So we had to have this done on our own account. Unfortunately this costs money. Raising questions, doubts and speculations does not. So the world is teeming with statements but lacking results and evidence. Finally we set out to put an end to this and spent the money for a measurement. Feel free to repeat our efforts, financed from another source but the copper industry! Feel free to provide us with measured results from other lamp types (instead of allegations) if you can afford more than we could! Afterwards let us continue discussing. But so far it is evident that with the configurations we have had tested so far the input energy saving was greater than the output energy loss. Nothing more did we allege. We do not doubt there are many other claims around but you must have read those elsewhere.

By the way, accounting for another request we recently received we underwent another expenditure and had the new 51 W T8 lamp by Philips tested, which is totally compatible with the present commonplace 58 W model. We will publish the results in the foreseeable future. So far we must say: Major loss of output for a minor saving on the input side when operated with a magnetic ballast, about proportional input savings versus output losses with an electronic ballast. What we unfortunately were not able to include is the influence of temperature, so feel free to support us with your own results!

Of course, by principle it does not take a fancy voltage reduction gadget to achieve the achievable saving. Not in principle. But we concentrate our efforts on solutions the existing user is able to buy on the market and implement. Note that we also point out that e. g. using a 240 V magnetic ballast in a luminaire rated 230 V or 220 V also does the job (provided you use electronic starters).

Also note that it says there the loss in a magnetic ballast is approximately proportional to the square of the current! So if you call our statement “a little simplistic”, then you are just quoting us. Of course it is! This is no science. The practician needs approximate and easily understandable data and implementable instructions. This is what we want to provide, not indigestible science. May I guess that your employer is a university?

Next, you say “A different way of achieving lower ballast losses would be to use a more robust ballast (thicker windings, etc.) to achieve lower losses at the same power.” Again, you are only quoting what we say (add “More and / or better magnetic steel” to make the list of the most important items complete and “simplistic” enough for the practicioner). This is nothing else but what is actually done, shifting from ballasts class D through C through B2 and hopefully one day up to B1 – but perhaps rather on a voluntary basis, for we see some drawbacks in a general prohibition of B2. So do the producers.

Of course the loss of light does most certainly become noticeable at some point. We do point this out, and more than this, even if not, we also claim that the user must be notified of this loss, even if it is noticeable only on a metering instrument! Anything else would be dishonest.

We are talking of energy savings here, i. e. active power. Please do not confuse with reactive power! This is dealt with very much in detail elsewhere (see http://lighting.copperwire.org/5.1.php). Not everything at once, please! The poor practician!

Of course the best energy efficiency will be achieved by appropriate choice of equipment to operate at the rated supplied voltage and not by adding additional equipment in front of sub-optimal appliances. Unfortunately not all appliances are optimized with respect to efficiency but rather to maximum output. In fact a fluorescent lamp with a magnetic ballast is at present the only device we can think of which overall works better at (slightly) reduced voltage. All others lose more at the end than is saved at the beginning, if any at all. Mind our publications on this.