Non-linear loads are divided into two categories - single phase and three phase loads. Each presents a potentially unique set of problems as a result of their characteristic harmonic spectrum.
Single-phase non-linear loads of the type found in servers used in this application generally contain a simple switch mode power supply. Switch mode power supplies produce a harmonic spectrum that is particularly rich in third harmonics, however other harmonics are characteristically present.
The negative effects of harmonics fall into two categories: those of current distortion related problems and those of voltage distortion problems. A brief summary of the effects of each is shown below:
Effects of high current distortion:
Overheating of conductors and insulation degradation
Neutral overload *
Increased transformer losses (need to over-size)
Nuisance tripping of circuit breakers
Neutral-earth potential *
Significant voltage distortion on networks with generators
Overheating and possible resonance with capacitors
Lighting ballast failures
PC monitor stroboscopic effect *
Mal-operation of microprocessor-based equipment *
Re-injection of harmonic currents into the utility network
Voltage Flat topping * Caused by Third (neutral current) harmonics flowing across network impedance
Table 2: Effects of high voltage distortion:
Causes linear devices to draw non-linear current
Torque pulsation in motors
Capacitor dielectric failure
Insulation breakdown Electronic lighting failure & Current Distortion Issues.
PC monitor and power supply failure
Both single and three phase non-linear (electronic) loads draw harmonic current. The dominant harmonic produced by single-phase loads is third, however its harmonic spectrum is rich in other harmonic currents (see figure 1, below).
Effects of Voltage distortion.
As summarised in table 2, voltage distortion causes negative effects on the entire distribution network not just those in the ‘path’ of the harmonic currents. Voltage distortion has two origins: Background distortion on the utility network and distortion generated on the customers’ network. Customer generated voltage distortion is normally well in excess of background (utility transmitted) voltage distortion and adds to it within the LV distribution network. Voltage distortion is relatively easy to understand. Ohms law states that a current flowing across impedance will generate a voltage drop. In the case of harmonic currents, this voltage drop will be at the frequency of the harmonic current causing it. In other words, as harmonic currents flow across the impedance of the distribution network, they cause a harmonic voltage drop, which adds to the fundamental voltage causing it to be distorted. As a common voltage links all loads, all equipment on the same bus will feel the effects of the distorted current. What this means is that loads which would otherwise draw linear current (current that is in linear proportion to the applied voltage) will now draw non-linear current in (linear proportion) to the non-linear voltage. In effect, non-linear voltage (Voltage distortion) makes the entire network non-linear, not just the electronic loads.
The magnitude of the voltage distortion is in direct proportion to the impedance of the network. In other words, as we increase the impedance of the network, the voltage distortion will increase.
Limits for Harmonic Distortion – G5/4-1
G5/4-1 set limits on the voltage distortion at the point of common coupling. (The point of common coupling being the point where the customer’s network connects to the utilities). Although historically reluctant to enforce it, Utilities are increasingly seeing higher voltage distortion on their networks and can, if they choose, enforce these limits. The spirit of the recommendation is one of reducing the risk of damage or malfunction to other consumers’ or the Electricity Board’s equipment. As a consequence of this our results will be compared to the recommendations for a 400V system, as a guide to best practice. The aim being to minimise the risk of damage or malfunction to equipment on the system under test.
Electricity Association Recommendation G5/4-1 Planning Levels for Harmonic Voltage Distortion and Connection of Non Linear Equipment to Transmission Systems and Distribution Networks in the UK.
+44 (0) 20 7706 5100 (Copies can be obtained from the Publications Department):www.energynetworks.org.uk
1. IEC 61000-2-4
The compatibility levels for individual harmonic components of the voltage shall be understood to relate to quasi-stationary or steady-state harmonics and are given as reference values for both long-term effects and very short-term effects. The long-term effects relate mainly to thermal effects on cables, transformers, motors, capacitors, etc. They arise from harmonic levels which are sustained for periods equal to or higher than 10 min.
With reference to long-term effects, compatibility levels for individual harmonic components of the voltage are given in Tables 2 to 4. The corresponding compatibility levels for the total harmonic distortion are given in Table 5.Very short-term effects relate mainly to disturbing effects on electronic devices that may be susceptible to harmonic levels sustained for 3 s or less. Transients are not included. With reference to very short-term effects in class 1 and class 3, compatibility levels for individual harmonic components of the voltage, and for total harmonic distortion, are 1,5 times the values given in Tables 2 to 5.
1. BS EN 50160 – Main requirements
EN 50160 gives the main voltage parameters and their permissible deviation ranges at the customer’s poin tof common coupling in public low voltage (LV) and medium voltage (MV) electricity distribution systems, under normal operating conditions. In this context, LV means that the phase to phase nominal rms voltage does not exceed 1000 V and MV means that the phase-to-phase nominal rms value is between 1 kV and 35 kV. The following table gives an overview of requirements to meet BS EN 50160
Flicker is caused by load switching within electronic apparatus and is commonly produced by devices such as arc-furnaces, rolling-mills and multiple welder loads, electronic ballasts, light dimmer switches. When the supply cannot fulfil the current demand, the ac voltage will temporarily dip and the effect on a 60W incandescent light bulb connected to the same supply point would be a temporary reduction in light, which if repeated would constitute flicker. The amplitude and frequency of these deviations can cause incandescent lamps to flicker. This is not only potentially annoying, but it can trigger seizures, especially in people with epilepsy.
How is it measured?
The measuring method is based upon measuring the RMS value of the voltage for each half-cycle and referring this to the average RMS value of the voltage. Not all variations are equally perceived by the human eye in terms of luminance change; therefore, these fluctuations are split into the frequency spectrum from 0.5 Hz to 25 Hz, with every frequency being individually weighted. The maximum weight is given for an 8.8 Hz voltage fluctuation. From weighted fluctuation values a new total RMS value is obtained, which, once measured over a period of 10 minutes, indicates the short term flicker severity, ‘Pst’
For evaluation of complex load fluctuations, a meter is used that simulates the EYE/BRAIN response (as defined in IEC 868). Using filters and statistical analysis, two values are produced. ‘Pst’ is the short-term flicker severity evaluated over a short period (10 minutes). ‘Plt’ is the long-term flicker severity evaluated over a long period (2 hours). From weighted fluctuation values a new total RMS value is obtained, which, once measured over a period of 10 minutes, indicates the short term flicker severity, ‘Pst’. The flicker is considered to be perceptible for Pst > 1. The evaluation of long-term flicker, ‘Plt’, is done from a sequence of 12 ‘Pst’ values over a two-hour interval.
BS EN50160:2000 (Voltage characteristics of electricity supplied by public distribution systems) states that for medium voltage supplies, a long-term flicker ‘Plt’ value of < 1 for 95% of the measuring period of one week should be attained.
Power factor correction equipment selection.
Today's factories, commercial buildings and homes contain more and more electronic loads. These loads, by nature of their energy efficiency, draw current in short bursts. Because this current is drawn in short peaks, it leads to the generation of Harmonics. Harmonics are currents which are integer multiples of the fundamental frequency. Harmonic currents cause additional heating in conductors, higher losses in magnetic devices such as motors and transformers and additional stress on electrical insulation and dielectrics.
As these currents flow across the network impedance, they cause a harmonic voltage drop which subtracts from the fundamental (50Hz) voltage causing it to be distorted as well. Examples of equipment which produces harmonic currents include: Variable speed drives, motor soft starters, robotic equipment, welders, uninterruptible power supplies (UPS), computers, fax machines and copiers, computer monitors, television sets and high frequency lighting like the type that is found in almost all modern commercial buildings.
Like all other loads, Power Factor Capacitors are also susceptible to the negative effects of harmonic currents and voltages. Because capacitors are naturally a low impedance path to harmonic currents, they will absorb a high percentage of the harmonic currents that are flowing in the distribution network. This leads to overheating of the capacitor elements and/or dielectric failure causing greatly reduced product life. In extreme cases, the capacitors can interact with the distribution transformer to create a resonant circuit which can actually cause an amplification of the harmonic currents in the network.
In order to ensure longevity of the product and years of trouble-free operation, the following guidelines are presented. These guidelines are based upon the percentage of harmonic generating load on the same transformer as the capacitors. To calculate the percentage of harmonic generating load on the transformer, add up the kVA of the harmonic generating loads and divide these by the total connected load. For example, if you found you had 250 kW of motors controlled by variable speed drives, plus 300kW of high frequency high-bay lighting, by rough estimate then you have 550 kW of non-linear load. If then, your total load was 2000 kVA on a transformer which was sized at 2500 kVA, you would have 550/2000 = 27% non-linear load.Using this value, you would select Detuned PFC from the description below:
Less than 20% non-linear load:
Standard automatic power factor capacitors are suitable. Because harmonic currents can be imported into your facility from the utility network, always specify 460V dielectric capacitors. These capacitors can absorb some harmonic current without damage. Note: If the amount of kVAr exceeds 0.33 x kW, detuned capacitors are recommended to avoid resonance with any harmonics which may be present on the network 20% to 50% non-linear load:
Detuned automatic power factor capacitors are suitable. Detuned capacitors employ reactors in series with each capacitor stage to limit the amount of harmonic current that is absorbed and to prevent resonance between the supply transformer and the newly installed capacitor bank. It is important however to remove any existing capacitors prior to installation of a detuned bank as their presence on the network can change the tuning characteristics of the detuned bank. An added benefit of a detuned bank is that it can absorb up to 50% of the harmonic current in the network which will reduce the overall voltage distortion.Over 50% non-linear load**:
PFC capacitors are essential in improving network efficiency and reducing CO2 emissions. However wrong application can lead to many related power quality issues which are avoidable. Obviously the cost of the different types of PFC vary quiet dramatically however it is usually a false economy to choose the cheaper option because the longevity of this equipment is substantially reduced.
Optimising or reducing the voltage has had mixed results which are very dependent on the types of loads applied but in theory reducing the voltage the kWh will reduce (P = V2/Z). From this formula a 2.5% reduction in voltage can lead to a 5% reduction in kWh. This is actually supported by many including the following statement; High voltages are actually having two negative effects; Reducing the life expectancy of the connected equipment i.e. a 230V rated lamp operating at 240V will achieve only 55% of its rated life” (failing after 550hrs instead of 1000 hours)*.
Increased energy consumption i.e. a 230V linear appliance used on a 240V supply will take 4.3% more current and consume almost 9% more energy*.
* Extracts from IEE 16th edition guide BS7671.
In many cases there is no disputing the obvious technical merits of reducing the voltage it is invariably the commercial cost versus payback on how it is achieved. By implementing the secondary tap down* scenario there are inherent risks of the supply voltage variation changing dramatically in the future as dictated by local network usage ie new large user will reduce the voltage on the network and the shut-down of a large end user will increase the voltage. However the initial costs are small and financial payback relatively quick. * Easily achievable providing regular maintenance of the transformer has been conducted.
In comparison the initial costs are high with the implementation of the optimisation solution but the voltage can be more regulated giving potentially greater kWh reduction (typically 8-12%) and not directly impacted by local network demands (NB. standard optimisation does not protect against sags on the network).