Wednesday, 14 May 2014

How to correctly determine the self-inductance Le of Power Capacitors

The self-inductance of power capacitors cannot be measured by means of an LCR-bridge. There are two reasons for that:


1. The tolerance band of any LCR bridge is too wide for a reliable measurement of the self-inductance of power electronics capacitors.

2. Whenever a capacitor is measured, the LCR bridge is connected with a series of a.) very large capacitance, b.) very small inductance, and c.) series resistance (i.e. sum of all conductors in the capacitor including terminals, wires, film coating, etc.) Both C and RS are representing such large and dominating values that it is practically impossible to get a reasonable reading of L. Example: The readings received by one client were in the range of 800…1000nH and more which is completely irregular and illogical.

The above is the reason why capacitor standard IEC 61071 contains detailed instructions in section 5.12. as to how the self-inductance of capacitors has to be determined. This is also an obligatory part of the type test procedure prescribed for power electronics capacitors and required under type test clause 5.2.2.j
The standard emphasizes that this can only be made indirectly through calculation from the resonance frequency, and that it is essential to measure the resonance frequency by the help of a procedure which excludes any mistakes or errors, caused by connections or auxiliary equipment. There are two different procedures.

Procedure 1: Frequency Run Method
Measuring equipment required: frequency generator, amplifier and high-definition circuit analyzer
By the help of the frequency generator, a defined frequency range is checked. At the same time the voltage change is read by a voltmeter. The frequency value at which the voltage value reaches its minimum defines the resonance point of the capacitor.
Afterwards, the self-inductance is calculated using the formula

Fres = 1/2pÖ(LC)

Procedure 2: Surge Discharge Method
A digital oscilloscope records the curve of the discharge during a surge discharge of the capacitor. Afterwards, the resonance frequency is identified by the numbers of intersection points over time scale.
It has to be noted that in method 2, there is an uncertainty factor which is the external inductance of the measuring circuit.

Hence procedure 1 is the most reliable method and therefore, we are recommending using this method only for the determination of the self-inductance of capacitors.

Also we would like to mention that selection of high quality measurement equipment , (Keithly for example) is key to the correct measurements.

As one can judge from the above, it is not reasonable to measure the self-inductance of capacitors as part of the routine test. This would require far too much manpower and time and cripple any cost-efficient production. This is also the reason why the applicable IEC standard 61071 does not list such measurement as part of the obligatory routine test for power electronics capacitors.

For more information contact us at www.electronicon-se.com 

Comparing Round Winding Technology with Flat Winding for Power Applications

It must be acknowledged that Flat winding technology is still being used in modern capacitors, in particular by

1.  many manufacturers of medium and high voltage ALLFILM capacitors for power factor correction (i.e. non-self-healing, oil-impregnated capacitors with aluminium electrodes and double layers of polypropylene dielectric), and by

2.  few manufacturers of power electronics capacitors (self-healing, metallized polypropylene film, dry dielectric)

It is claimed by the manufacturers offering flat winding technology that these capacitors can be smaller and, as a consequence, save material cost. In reality, the flat winding technology bears serious technical and mechanical disadvantages if compared with cylindrical windings. These disadvantages may become very essential when considering capacitors for long operating periods as are required, for example, in HVDC and traction applications.

Technical Aspects:

    Initially, a flat capacitor element is wound very similar to round elements; however this is made on a core with large diameter which is removed after the winding process. The element is then pressed flat to form a stable, quite solid block.

The main technical flaw of this technique lies in the pressing itself: Even though flattened windings which are produced on top-modern equipment look absolutely homogenous at first glance, this smooth impression disappears as soon as the windings are opened and unwound. By the laws of geometry, the outer film layers on the winding must get stretched whilst the inner layers get compressed and wrinkled; moreover, the innermost winding turns get folded with a sharp edge. It is obvious that the stretching of the outer film layers does not contribute to their long-term stability. But even worse is the fact that, as opposed to the perfectly uniform electrical field in round windings, the wrinkles and the sharp folding edges in the flat winding cause undefined conditions of the electrical field; the field strength can nearly double in the areas of the bending edges! It is logical that this must contribute to a higher failure risk.


One of the most important parameters for the design of polypropylene film capacitors is a high and homogenous pressure among the film layers. High pressure ensures the hermetical closure of the winding element which is a very critical pre-condition for long life-time and stable behaviour during the entire operating period.Thanks to their compact round shape and the defined shrink of the film during the thermal  treatment (which is an obligatory part of our production process), round capacitor windings are self stabilizing perfectly. Their internal pressure remains constant throughout the operating time, and the windings cannot change their volume or shape anymore. 
     
    Flattened windings, however, need to be pressed externally, not only during the production process but also during the entire operating life. It is clear that the external pressure which can be implemented on the winding is limited by the capacitor construction, and that it is impossible to maintain this pressure uniform and unchanged during the entire operating period.

    Polypropylene elements are contact-sprayed (“schooped”) with a zinc layer at both ends in order to establish electrical contact with the metallization on the film. Whilst cylindrical windings are covered completely and uniformly at both ends by this layer, the flat windings cannot be contact-sprayed in the area of the sharp bend which is – as shown above - the most critical area of the element anyway. As a consequence, the current for energizing these folded areas has to enter the winding element through the parts of the shooping layer which are in the vicinity of the bending, causing higher current density (and potential for local hotspots) in these areas which increase the risk of failures.

Moreover, the uniform geometric condition of round-shaped windings allows for an optimum penetration of the zinc-particles during the shooping process (assuming proper know-how of the manufacturer, of course), for perfect contact with minimized losses.
Such optimization is unthinkable with flattened windings which bear a higher risk of poor contact which may result in more local hotspots.

 Anybody who has seen a rectangular reactor winding after heavy surge charges, knows about the power of surge currents: under heavy current stress, any winding, no matter if forced into flat or rectangular shape, strives towards its optimum shape, which is round. During heavy surge discharges (e.g. external fault situation causing short circuit discharge of the capacitor), the uneven current distribution inside the flat windings produces additional mechanical stress within the elements (also compare pt. 2) which may deteriorate or damage the link between schooping layer and metallized film. This will have a negative effect on the operating life (or failure risk) of the capacitor. Obviously, round windings have no such problem.

Flat windings can only be efficiently displayed if (a)  the outer housing size fits exactly the width of the (flat-pressed) winding (b) the film has a considerable width (otherwise, short flat-packs will require a higher number of windings to be assembled which increases cost).

(a.) will pre-define the size of these capacitors. They can only be altered in certain size steps because they must make efficient use of the flat packs. A manufacturer of flat-winding capacitors will therefore always strive to adjust the customer’s specification to his “ideal” sizes in order to achieve the optimum usage of space for his flat windings and make it more difficult for competitors to match. This explains why such manufacturers often try to “buy” the business by a very low initial price. This may become a cost trap later when the customer depends completely on these dimensions (and on the supplier).
Even though it is true that round windings do not fill a rectangular space as completely as flat windings, manufacturers with round windings are more flexible in adapting their shape and size to the customer’s project requirements.

(b.) means a long current path inside the winding which always contributes to higher series resistance Rs. The series resistance, however, is responsible for the current power losses of a capacitor; hence it contributes to the temperature rise inside the capacitor and, as a consequence, to the failure rate. With round windings, the manufacturer is free to optimize the winding element depending on current, capacitance, ambient temperature conditions, and available space. Of course, the length of flattened windings can be reduced as well in order to improve the series resistance. However, since the diameter of the original round bobbin for the flat winding cannot be increased at will, any reduction of the length of flat-packs must in turn be compensated by a bigger number of windings to be assembled. This would then increase cost.

From a technical and commercial point of view the flat windings do not offer any significant advantage, it is therefore recommended not to disturb the geometry and stay focussed on the round windings.







Sunday, 19 January 2014

Detuned Reactors - Adjusted and Non-adjusted Calculations


In this section I shall try to give instructions to select a detuned reactor (reactor -capacitor combination)

The table below is given in format which allows reader to print and use it as it is.

Steps
Fill Below
Remarks/Actions
1.     Network voltage


2.     Frequency

For 60Hz. Contact us directly.
3.    Harmonic spectrum is known


3rd

14% detuning if 3rd present
5th

7% to 5.6% if 5th present
7th


9th


4.   Harmonic currents are known

Go to 5th step
 Ih > 1.5 In

Consult Specialist
1.3In > Ih > I.5In

5.67%
1.1In > Ih > I.3In

7%
5.     Resonant conditions exists

Select 7%
6.   No info on Harmonic currents

Select 7%
7.     Select compensation steps


8.     Adjusted or non-adjusted

See example calculations below

9.     Reactor material

Recommend Aluminium
10.   Linearity

Standard or customised.
11.   Terminations


12.   Thermal cutouts

Recommended
Selection complete. Refer PFC components Catalogue for part numbers.



25 kvar 440 V 50Hz 3 x 137uF (411uF)
To be detuned to 189Hz (7%)

Selection of the reactor-





25 kvar 440 V 50Hz 3 x 137uF (411uF)
To be detuned to 189Hz (7%)


Using the above methods it is easily possible to arrive at the correct values for the detuned combination. 

For more information please visit Electronicon-SE or contact us directly.

Friday, 17 January 2014

Understanding Detuning Reactors ( Anti-resonance Reactors)

Reactor protection

Most automatic capacitor banks employed today are provided with reactor protection as a result of the in­creasing harmonic loading of the consumer installation and the power networks. Every capacitor or capacitor tap is connected in series to an inductance (reactor), in contrast to "normal" unprotected compensation.

If the resonant frequency of the series resonant circuit formed in this way ( capacitors and Inductor) deviates (is lower) by more than 10% from the frequency of the nearest harmonic, then one speaks of a detuned resonator circuit or an anti-resonance circuit. Reactor protected compensation systems are designed as detuned resonator circuits and the series resonant frequency f0 is normally chosen to be below the fre­quency of the 5th harmonic (250 Hz). The capacitor and reactor system is therefore inductive for all harmonic frequencies ³ 250 Hz and dangerous resonance between the capacitor and network inductance (e.g. transformer) is therefore avoided.  Consumer installations with high 3rd harmonic (150 Hz) components are an exception but it can become necessary to set the series resonant frequency to 134 Hz in such cases.


The graph above shows the frequency response of 5.67%; 7% and 14% detuned circuits. It should be noted that the closer the resonant frequency of the anti-resonance filter is to the Harmonic to be filtered, lower is the impedance offered and therefore better is the filtering effect. For example 5.67% which is tuned at 210Hz will have lower impedance to the 5th harmonic compared to the 7% detuned which has a resonance frequency of 189Hz. This is shown by the points on each curve at 250Hz. Below table shows the amount of harmonic current that will be absorbed in each case. 

Detuning
Resonant Frequency
% of 5th Harmonic current absorbed
5.67%
210 Hz
30~50%
7%
189 Hz
10-20%
14%
134 Hz
0%

From the above table it is clear that depending on the load profile and also the duty cycle of the non-linear loads the proper selection of the detuning frequency should be made. Please consider that using 7% detuned system may not be always helpful and high distortion may still persist on the load busbars.

If the series resonant frequency is between 10% below or above a harmonic frequency, then it is de­scribed as a tuned resonator circuit. Tuned resonator circuits are normally employed as wave traps for the deliberate reduction of individual harmonics.

Reactor protection-factor p

The reactor-protection factor p [%] specifies the ratio of the reactor reactance to the capacitor reac­tance at network frequency.

p=XLx 100 / Xc

The resonant frequency of the series resonant circuit can also be calculated from p using the following equation:

fres = f1 x (1/sqrt(p)) 


For example:  p=7 %, f1 = 50 Hz  
fres = 50 x (1/sqrt(0.07)) = 189Hz


One of the often-tried standard values is normally used for the choice of a suitable reactor-protection factor for the application:

Reactor-protection factor
p
Series resonant
frequency fR
5%
223 Hz
5.67%
210 Hz
7%
189 Hz
8%
177 Hz
12.5%
141 Hz
14%
134 Hz


Capacitor rated voltage with reactor protection

A voltage increase arises at the capacitor from the serial connection of the reactor and capacitor. It can be calculated from the reactor-protection factor p:
For example: p = 7%, UN = 440 V
Uc =Un(1/(1-p))

Uc= 400(1/(1-0.07)) = 473V

The capacitors employed for p = 7% must therefore be suitable for a continuous rated voltage of at least 480 V. Here, you must always be careful about the voltage tolerance for the nominal net voltage.


When the voltage on the capacitors increase the KVAr output of the capacitor bank also changes. This is given by the following equation.

Qc=(1-p/100).[(Uc^2)/ (Un^2)]   .Qn

In such a case case it is necessary to use adjustable ratings. For more information please read the reactor application note from Electronicon System Electric . or read the entry for calculations.