BS 5049-1:1994 download.Radio interference characteristics of overhead power lines and high-voltage equipment Part 1: Description of phenomena.
I Scope
This publication applies to radio noise from overhead power lines and high-voltage equipment which may cause interference to radio reception, excluding the fields from power line carrier signals.
The frequency range covered 18 0.15 MHz to 300 MHz.
2 Object
This publication has been prepared in order to provide information on the many factors involved in protecting the reception of radio and television broadcasting from interference due to high voltage overhead power tines and associated equipment.. The information given should be of assistance when means of avoiding or abating radio noise are being considered. The publication does not deal with interference due to the fields produced by power line carrier signals.
Information is mainly given on the generation and characteristics of radio noise from a.c. power lines and equipment operating at 1 kV and above, in the frequency ranges 0.15 MHz to 30 MHz (a.m. sound broadcasting) and 30 MHz to 300 MHz (f.m. sound broadcasting and television). The special aspect of spark discharges due to bad contacts is taken into account.. Some information is also given on interference due to d.c. overhead lines for which corona and interference conditions are different from those of s.c. power lines. Interference due to the current collecting equipment of overhead railway traction systems is not included.
The general procedure for establishing the limits of the radio noise field from the power lines and equipment is given, together with typical values as examples, and methods of measurement. The clause on limits concentrates on the low frequency and medium frequency bands as it is only in these that ample evidence, based on established practice, is available. No examples of limits to protect reception in the frequency band 30 MHz to 300 MHz have been given, as measuring methods and certain other aspects of the problems in this band have not yet been fully resolved. Site measurements and service experience have shown that levels of noise from power lines at frequencies higher than 300 MHz are so low that interference is unlikely to he caused to television reception, The values of limits given as examples are calculated to provide a reasonable degree of protection to the reception of broadcasting at the edges of the recognized service areas of the appropriate transmitters in the a.m. radio frequency bands, in the least favourable conditions likely to be generally encountered. These limits are intended to provide guidance at the planning stage of the line and standards against which the performance of the line may be checked after construction and during its useful life.
Recommendations are made on the design, routing, construction and maintenance of the lines and equipment to minimize interference and it is hoped that this publication will aid other radio services in the consideration of the problems of interference.
Other investigators believe that, if a quasi-peak detector is used to measure the field strength, the individual pulses do not add and others have obtained results between the two extremes. This disagreement is only important in analytical prediction methods, the results obtained by the different methods vary by only 1 dB or 2 dB.
In the case of multi-phase lines, the calculation follows the same principle but is complicated by the presence of several modes, each mode having a different attenuation coefficient. A more detailed discussion, with examples of calculation, is given in Clause 6.
4.1.2 Definition of noise
The instantaneous value of the noise varies continuously and in a random manner. but its average power level over a sufficiently long period, for example. 1 a. gives a stationary random quantity which can be measured. Another quantity suitable for measurement is the peak or some weighted peak value of the noise level.
A noise measuring set is basically a tuneable selective and sensitive voltmeter with a specified pass-band. When connected to a suitable rod or loop aerial and properly calibrated, it can measure the electric or magnetic component of the noise field.
Depending on the design of the measuring apparatus, the noise level can be measured in terms of r.ms., peak or quasi-peak values. The r.ms. value defines the noise in terms of energy. Many types of noise from electrical equipment, as well as noise due to power-line corona, consist of a succession of short pulses with approximately stable repetition frequencies. In such cases the nuisance effect of the noise can be realistically indicated by a quasi-peak type of voltmeter rather than by the r.rns. type. The quasi-peak value is obtained from a circuit which includes a diode and a capacitor with relatively short charge and long discharge time constants. The voltage on the capacitor floats at a value somewhat below the peak value and depends on the repetition rate, that i8 to say a weighting feature is included in the response. This principle is adopted in the C.I.S.P.R. measuring instruments, details of which are given in C.I.S.P.R. Publication 16, The noise level is thus defined by the value measured by such an instrument expressed in microvolts or microvolts per metre. Using the ratio of the electric to magnetic fields, K/H = 377, the measured values can be expressed by convention in microvolts per metre even for instruments using a loop aerial responding to the magnetic field.
4.1.3 Influence of external parameters
To determine the corona inception gradient of a cylindrical conductor with smooth surface. Peek’s formula is often used:
Atmospheric conditions likewise play an important part, In conditions of rain, fog, snow or dew, drops of water form on the surface of the conductor and at low temperatures ice can form. This further reduces the corona inception voltage and increases the noise level as shown in Clauses 5 and 6.
With regard to bad contacts and the production of small sparks, the effect of rain and humidity is to bridge the relevant gaps either by water droplets or by humid layers, thus reducing the level of this type of noise. Rain and humidity thus affect corona noise from conductors in a way opposite to that due to had contacts. Hence when interference is observed during rain or fog, it can be concluded that it is caused by corona. On the other hand, when interference is observed during fine weather and disappears or decreases during rain or fog, it is due to bad contacts.
4.2 Main characteristics of the noise field resulting from conductor corona
To rationalize the measurement of radio noise from a transmission line and facilitate comparisons between different lines, it is desirable to standardize the conditions under which the measurement is to be carried out.
The main characteristics of the noise field are the frequency spectrum. the lateral profile and the statistical variation of the noise with weather conditions. It is assumed as a first approximation that these
characteristics are independent of each other.
4.2.1 Spectrum
The spectrum is the variation of the radio noise measured at a given point in the vicinity of a line, as a function of the measurement frequency. Two phenomena arc involved:
a) Current pulses
The current pulses generated in the conductors by the discharges show a particular spectrum
dependent on the pulse shape. For this type of discharge the measured noise level falls with frequency.
In the range of broadcasting frequencies, where the positive discharges have a predominant effect, the spectrum is independent of the conductor diameter.
b) AUenuat ion
The attenuation of noise propagating along the line increases with frequency. This efTect modifies the spectrum by reducing still further the noise level with increase in frequency.
The measured spectra are often fairly irregular because of the standing waves caused by discontinuities such as angle or terminal towers or abrupt ground level variations. In addition, the noise generation might vary whilst the measurements are being made.
To aid prediction calculations, “standard spectra” are used. Experience has shown that all spectra can be put into two families, one applying to horizontal conductor configurations, the other to double.circuit and triangular or vertical conductor configurations. The difference between these two families originates from the phenomenon mentioned in Item b) above, the propagation differing slightly according to the type of line.
6.2.2 Analytical methods
The correlation between noise current and noise field can also be evaluated by means of analytical methods similar to those already described in the case of the corona effect on conductors (see Sub-clause 5.3). When the radio noise current I, generated by an individual source on a conductor and injected into the conductor, is known, the determination of the radio noise field F. produced at a given position with respect to the conductor, is carried out by considering, in the first place, the splitting of the current I between the two sections of line, as seen from the injection point. For example, in the case of a source of noise on a line of infinite length, the current is divided equally between the two sections of the line. The attenuation of the current propagating along the conductor is then calculated and, finally, the field produced by the current at a given position is evaluated.
In the case of lines with only one conductor, for example a monopolar d.c. line, the calculation process is relatively simple, as all it calls for is a knowledge of the attenuation constant as a function of the frequency and the resistivity of the soil.
In the case of lines with more than one conductor, three-phase a.c. lines, bipolar or homopolar d.c. lines, the calculation of propagation of the noise is less simple and is generally dealt with by modal analysis, The complete modal theory is relatively complex and various more or less simplified procedures have been developed L2, , 42. 43. 44]. The principle, however, remains substantially the same and the actual system of radio noise currents, or voltages, is reduced to a few simple systems, characterized by simpler laws of propagation similar to those that exist for a system with only one conductor. It is then a question of applying similar calculations to each system and then aggregating the individual fields in order to determine the resultant field.
Where several sources are distributed on one of the three phases, the calculation process is much the same as previously described for the single source. In this case, account has only to he taken of the aggregation of the various sources of noise which are usually assumed to be of the random type.
In the case of noise sources on all three phases, the calculation of the field is carried out separately for the noise injected into each phase, and the total field E is obtained by the same processes as those described in Sub-clause 6.2.1.1.
6.2.3 Example of application
An example using the analytical method described above has been worked out with reference to a 420 kV line of infinite length having an average span length of 400 m and insulator strings producing a radio noise voltage, when referred to 300 ft of 49.5 dB above 1 ‘iV. that is, a current of 1 pA per string.
This radio noise may be emitted directly from the valves and associated equipment comprising, in this instance, mainly the feeders and the bushars of the convertor station. These busbars will often be of considerable length and well able to act as efficient radiators. The convertor will be. of course, connected to incoming and outgoing s.c. and d.c. circuits and these may both be of overhead line construction. The radio noise will be guided and emitted from such overhead lines
Vithout any suppression measures, the radio noise level could be intolerable and it is, therefore, necessary to reduce this level to an acceptable value. This can be achieved by different methods which depend on the valve type and the technical installation of the substation.
In most of the earlier operational high voltage d.c. schemes, mercury valves were used. The technical layout of this valve type requires a hail which protects the valves against environmental influences and makes it possible to keep temperature limits within a specified range. By electromagnetic screening of these halls the radio noise level outside can be reduced considerably. Using solid metal sheets or wire mesh, an attenuation value of 40 dB to 50 dB for frequencies between 0.15 MHz and 5 MHz is possible. To reduce the noise passing through the valve hail bushings, filters should be installed in all outgoing lines and the filtering of the d.c. lines must be especially effective. The convertor transformers, between the valve group and the a.c. lines, and the filter circuits on the ac. side. may reduce the possible conduction of radio noise from the convertor station to these a.c. lines.
In the case of thyristor valves the radio noise problem may be somewhat Less severe. Thyristor valves, when fired. may have a voltage collapse time of up to 25 is, compared with 1 .ts for mercury-arc valves. One reason for this is the use of damping circuits within the thyristor valve. With such a long voltage collapse time, it may not be necessary to screen the valve, ball, and radio noise filters may be simpler or even unnecessary.
A possible solution for reducing direct radiation from a valve is the installation of the thyristor valve within a steel tank which acts as an effective screen. Care should be taken that the tank does not radiate as a dipole or resonator.
The further development of high voltage d.c. convertor stations will tend towards more compact layouts with shorter connections between valve groups, convertor transformers and substations. This will have a favourable influence on the reduction of radio noise from convertor stations.
Generally, it is possible to reduce the radio noise produced by convertor valves and their auxiliaries to an acceptable level. The costs of such a reduction largely depend on the valve type and the convertor station design.