Flicker: Causes, Symptoms, and Cures

Theoretically, for any supply line, the voltage at the load end is different from that at the source. We can demonstrate this with a per-phase equivalent circuit, as shown in Fig. 2. Here, E is the source voltage, RS is the equivalent line resistance, X_S is the equivalent line reactance, Z_S is the equivalent line impedance, I_O is the current, and U_O is the voltage at the load terminals.

Depending on its cause, a voltage change can take the form of a voltage drop having a constant value over a long time interval, a slow or rapid voltage change, or a voltage fluctuation.

Voltage fluctuation is defined as a series of rms voltage changes or a cyclic variation of the voltage waveform envelope. The defining characteristics of voltage fluctuations are:

• The amplitude of voltage change (difference of maximum and minimum rms or peak voltage value occurring during the disturbance);

• The number of voltage changes over a specified unit of time; and

• The consequential effects (such as flicker) of voltage changes associated with the disturbances.

Until recently, voltage fluctuations in power systems, and at the load terminals, were characterized using factors associated with the peak-to-peak rms voltage change in the power system. The energy of voltage fluctuations, their power spectrum (also called the energy spectrum of voltage fluctuations), and their duration were taken into account when assessing voltage fluctuations. Currently, the basic parameters that determine voltage fluctuations are short-term flicker severity, called the P_ST index, and long-term flicker severity, called the P_LT index. (See sidebar “Flicker Levels and International Standards” on page 32.) These parameters refer to voltage fluctuation effects on lighting and their influence on humans.

Sources of voltage fluctuations. The primary cause of voltage changes is the time variability of the reactive power component of fluctuating loads. Such loads include arc furnaces, rolling mill drives, and main winders — all of which are loads with a high rate of change of power with respect to the short-circuit capacity at the point of common coupling (PCC).

Small power loads, such as starting of induction motors, welders, boilers, power regulators, electric saws and hammers, pumps and compressors, cranes, and elevators also can be sources of flicker.

Other causes are capacitor switching and on-load transformer tap changers, which can change the inductive component of the source impedance. Variations in generation capacity of wind turbines, for example, also can have an effect. Sometimes, voltage fluctuations are caused by low-frequency voltage inter-harmonics.

Mitigation of voltage fluctuations in power systems. The effects of voltage fluctuations depend first on their amplitude, which is influenced by the characteristics of the power system, and second, on the rate of their occurrence, which is determined by the type of load and character of its operation. Usually, mitigation measures focus on limiting the amplitude of the voltage fluctuations. The technological process is seldom influenced.

Examples of mitigation methods for various types of equipment include:

• Arc furnaces — Incorporate series reactors (or variable saturation); ensure proper functioning of the electrode control system; segregate and provide preliminary heating of charge.

• Welding plants — Supply the plant from a dedicated transformer; connect single-phase welders to a 3-phase network for balanced load distribution between phases; connect single-phase welding machines to different phases from those powering lighting equipment.

• Adjustable speed drives — Use soft-start devices.

Another way to reduce the amplitude of voltage fluctuations is to increase the short-circuit power, with respect to the load power, at the PCC to which a fluctuating load is connected. This can be done by:

• Connecting the load at a higher nominal voltage level;

• Supplying this category of loads from dedicated lines;

• Separating supplies to fluctuating loads from steady loads by using separate windings of a three-winding transformer;

• Increasing the rated power of the transformer supplying the fluctuating load; or

• Installing series capacitors.

Voltage stabilization solutions. Yet another way to reduce the amplitude of voltage fluctuations is to reduce the changes of reactive power in the supply system. You can do this by installing dynamic voltage stabilizers. Their effectiveness depends mainly on their rated power and speed of reaction.

By drawing reactive power at the fundamental frequency, dynamic voltage stabilizers produce voltage drops on the supply network impedances. Depending on whether the reactive power is inductive or capacitive, the rms voltage value at the PCC can be increased or reduced.

Figure 3 shows the classification of various solutions for dynamic voltage stabilizers. They are mainly 3-phase systems, of high rated power, designed for voltage stabilization at the main point of a distribution system or of a specific load or group of loads at a PCC.

Synchronous machines. These are traditional sources of fundamental harmonic reactive power (lagging or leading) supplied in a continuous manner. They also can be the source of mechanical energy when operated as a compensator and a motor.

The use of a synchronous machine with no excitation current control is pointless because, in order to reach the standard limit level of voltage changes, the machine would need a power rating several times greater than the power of the load requiring stabilization. This fact, as well as the required dynamic parameters of the stabilization process, requires the synchronous machine to be operated in a closed-loop voltage control system with fast excitation current control, as shown in Fig. 4. Such a solution enables a fast rise time of the machine's reactive current.

Static compensators. These devices (other than STATCOM) employ capacitive and/or inductive passive components that are switched, phase controlled or combined with controlled core saturation. They supply the required stabilizing reactive current either in discrete steps or, more often, in a continuously variable fashion. Static compensators are considered to be the most advantageous solution for improving the power supply quality, in both technical and economic terms.

Compensators with saturable reactors. These devices employ magnetic circuit saturation for voltage stabilization. Two of these solutions have found a wide practical application: self-saturable reactors (SRs) and reactors with DC control circuit. SRs were one of the first static compensators applied on an industrial scale to mitigate the effects of voltage fluctuations. They are designed so that, at the minimum of the voltage range, the core is just below saturation and a magnetizing current flows, similar to that of an unloaded transformer. In this state, it has practically no influence on the voltage magnitude. At nominal voltage, the reactor is saturated, so a small change in the supply voltage effects a considerable change in the current. The compensator is usually connected to the supply network without a step-down transformer.

Reactors with DC control circuits are most often operated with a parallel capacitor bank, which forms a filter for high order harmonics. It essentially works as a transductor, where adjusting the DC magnetizing current controls the primary current magnitude. The control DC winding is usually supplied from a fully controlled thyristor converter — the power of which does not normally exceed 1% of the stabilizer rated power. This solution enables the forcing of transient current, thus providing faster operation of the system. By adjusting the magnetizing current, the reactor's primary current changes from practically zero (unsaturated core) to the maximum value (saturated core) over the entire range of the primary current changes. A considerable disadvantage of this solution is the generation of the high order current harmonics. In the 3-phase version, a larger number of slots and appropriate linking of numerous windings allow the high order current harmonics to be practically eliminated, but at the cost of slower system response. The use of three single-phase stabilizers allows for correction of unbalance.

Thyristor switched capacitors (TSC). In this solution, the sectioned capacitor banks are connected phase-to-phase with each section switched (on or off) by means of AC thyristor switches (Fig. 5). Therefore, the values of the compensator equivalent susceptances change in a discrete manner, depending on the number of active sections. By providing a suitably large number of small sections, the required resolution of change of susceptance for a single step can be obtained. Synchronization of switching and initial precharging of the capacitors avoids the overcurrents and overvoltages normally associated with capacitor switching. Time of reaction for symmetrical operation does not exceed 20 milliseconds.

FC/TCR compensator. This solution is an example of indirect compensation. Depending on the required function (voltage stabilizer or reactive power compensator), the value of the sum of two components of the current is controlled. For example, to control the fundamental harmonic of the capacitor current, the capacitor is operated as a filter or as switched capacitor steps (TCR/TSC). For control of the fundamental harmonic of the reactor current, a thyristor AC switch is used.

Self-commutated converter voltage sources and reactive current/power sources. The compensator comprises a voltage source converter (VSC). The switching states of semiconductor devices (pulse width modulation) determine the value and character of reactive power (inductive or capacitive).

The most commonly used compensator is STATCOM, which is a new generation of static compensators that employs semiconductor devices with forced commutation. Its name — static synchronous compensator — is derived from the principle of operation, which is analogous to the operation of the synchronous compensator. The basic part of the compensator is an AC/DC converter, which is connected to the network via an inductive reactance, usually the leakage inductance of transformer. When the converter voltage is lower than the supply network voltage, the compensator is an inductive load. Conversely, when the converter voltage is greater than the supply voltage, the compensator supplies reactive power to the network, thus behaving as a capacitive load.

Big picture. Flicker is a subjective phenomenon. Consequently, it's difficult to determine the direct cost of its effect. Nevertheless, the phenomenon affects the ability to provide lighting that is steady and consistent. Certainly, it can affect productivity in an office or factory, but the cost of flicker usually is based on the cost of mitigating it when the complaints become significant.

Developments in power electronics, in particular in semiconductor device manufacturing, has enabled the practical realization of voltage dynamic stabilization systems of larger and larger rated power, while at the same time minimizing investment and operational costs. The availability of equipment with the ability to execute complex control algorithms also allows the use of diverse functions, including dynamic voltage stabilization.

Sidebar: Tracking Human Perceptibility of Flicker

Research on the process of visual perception has a history going back more than 40 years. Initially, it consisted mainly of tests carried out on selected representative groups of individuals using diverse light sources and various waveforms of voltage changes. On that basis, the researchers determined perceptibility and flicker severity curves.

These curves present values of sinusoidal or rectangular voltage fluctuations (vertical / Y-axis), and frequency (horizontal / X-axis). The area above the curve defines voltage fluctuations that produce noticeable, unacceptable flicker, whereas the area below the curve defines acceptable flicker levels.

Participation of physiologists and psychologists in these experiments allowed the development of improved mathematical models for the neuro-physiological processes. The resulting experiments offered the first opportunity to advance the thesis of similarity between the sensitivity of the human eye to light stimuli and the frequency characteristic of an electrical analog signal. Further studies took into account not only the amplitude of changes, but also different levels of eye adaptation to the average luminance.

Other studies demonstrated that the response of the human eye has the characteristic of a band-pass filter between 0.5 Hz and 35 Hz, with maximum sensitivity to the luminous flux at a frequency around 8 Hz to 9 Hz.

For incandescent light sources, voltage fluctuations of around 0.3% of the average value are detected at this frequency. Physiological effects depend on the amplitude of luminous flux changes, the frequency spectrum, and the disturbance duration.

The brain response to the light stimulus has an inertial characteristic with a time constant of about 300 milliseconds, meaning that slow changes of luminous flux are followed (noticed) and fast changes are smoothed (unnoticed). For instance, two short changes in the luminous flux, occurring within 300 milliseconds, are perceived as a single change. Short changes of luminous flux, followed by a longer pause, are more annoying.

The phenomenon of flicker is more dominant in the periphery of the visual field than in those areas on which the observer's attention is focused. The voltage fluctuation necessary to produce perceptible flicker is independent of the type of supply voltage (AC or DC) used for the lamp.

Sidebar: Flicker Levels and International Standards

According to Alex McEachern of Power Standards Lab, Alameda, Calif., flicker levels are characterized by two parameters:

• P_st: a value measured over 10 minutes that characterizes the likelihood that the voltage fluctuations would result in perceptible light flicker. A value of 1.0 is designed to represent the level that 50% of people would perceive flicker in a 60W incandescent bulb.

• P_lt: a value derived from 2 hours of P_st values (12 values combined in cubic relationship).

IEC 61000-2-2 specifies the following flicker compatibility levels:

• Compatibility level for short-term flicker (P_st) is 1.0.
• Compatibility level for long-term flicker (P_lt) is 0.8.

McEachern also says it's not always possible to maintain flicker levels within these compatibility levels. To address this shortcoming, EN 50160, “Voltage Characteristics of Electricity Supplied by Public Distribution Systems,” specifies less restrictive requirements for the supply system performance. The EN 50160 limit is that 95% of the long-term flicker values (P_lt) should be less than 1.0 in a one-week measurement period.

Note that individual step changes in the voltage, such as those caused by motor starting or capacitor bank switching, often are limited separately from the continuous flicker limits. IEC 61000, “Electromagnetic Compatibility (EMC),” in Part 2-2, “Environment - Compatibility Levels for Low-Frequency Conducted Disturbances and Signaling in Public Low-Voltage Power Supply Systems,” specifies a compatibility level of 3% for the individual voltage variations.

EN 50160 specifies a limit of 5% for these variations but mentions that more significant variations (up to 10%) can occur for some switching events. Specific recommendations are not provided in IEEE, but individual utilities usually have their own guidelines in the range between 4% and 7%.

IEEE also is adopting this method of characterizing flicker in IEEE 1453-2004, “Recommended Practice for Measurement and Limits of Voltage Flicker on AC Power Systems.”