Since the start of power transmission it has been found that pollution performance of insulators can severely impact system reliability. Insulators installed close to sources of contamination such as seacoasts, industrial and agricultural areas, highways, etc. can become covered with a conductive surface layer during wetting conditions. The result could be discharge activity, known as dry band arcing that, in severe cases, leads to flashover.
This problem led to considerable research and ultimately resulted in publication of important review documents. In 1986, after much discussion the IEC published its first guide for insulator selection with respect to pollution – IEC 815. With this document, the now well-known concept of creepage distance became standardized and, due to its simplicity, this concept was widely adopted. But its application was not always successful, especially in highly polluted environments where it was found that supplying enough creepage distance was not sufficient to guarantee good insulator performance. Further investigations revealed that a number of other parameters, such as profile and material, also needed to be taken into account when selecting and dimensioning insulators.
By the early 1990s it had become clear that IEC 815 needed substantial revision to address these shortcomings and also to accommodate new insulator technologies. A large-scale project was started to revise and update this standard to cover both AC and DC as well as ceramic, glass and polymeric insulators. This revision relied heavily on Cigré working groups to provide a firm technical basis.
Work started in 1994 within Cigré SC 33 with establishment of a reference document that could form the basis for future guidelines. This “review of current knowledge”, published as Cigré Brochure 158 in June 2000. The next step saw development of guidelines for the selection and dimensioning of outdoor insulators for AC systems, published as Cigré Brochure 361 in June 2008. The final task was to develop similar guidelines for DC systems, finally published in December 2012 as Cigré Brochure 518.
The aim of this article, contributed by Chris Engelbrecht, is to introduce this brochure to a wider public and to present the basic principles established by Cigré for selecting and dimensioning insulators for DC systems.
Differences Between AC & DC Pollution Flashover
There are a number of important differences between development of pollution flashovers under AC and DC energization.
1. DC energized insulators typically accumulate more pollution than AC energized ones. Under low wind speed conditions, the electrostatic attraction of pollution particles under the unidirectional DC electric field overrules deposition by aerodynamic action. On AC energized insulation, there is little to no attraction of pollution by the alternating electric field. Field measurements indicate that the ratio of DC to AC pollution deposition can vary from 1 to as much as 10.
2. With the absence of voltage zeros, dry band arcing under DC is more likely to grow into flashover than would be the case under AC, where dry band arcs need to re-ignite after each voltage zero. DC dry band arcs are also more mobile and more likely to leave the insulator surface to propagate through air. This necessitates development of special DC profiles to ensure effectiveness of the creepage distance. Another consequence of this difference in flashover development is the fact that the relative flashover strength of DC insulation deteriorates more than for AC insulation for any given increase in the pollution severity (illustrated conceptually in Fig. 1).
3. Relative to other stresses, pollution performance is often the parameter that dictates the insulation design of HVDC systems. This is different from AC systems where insulation distances are typically determined by required lightning or switching performance of the line or substation. This is conceptually illustrated in Fig. 2 for EVH and UHV system voltages and showing that, in polluted areas, extreme insulation lengths may be necessary in certain cases. This could force system planners to reconsider the whole conceptual design of a project to evaluate alternatives such as different line routes, to avoid severely polluted areas, or implementing cables or indoor switchyards to minimize the number of external insulation surfaces.
There is thus a clear need to consider the insulation design for HVDC systems already at a very early stage in a project and certainly much earlier than customary for AC systems. Also, it is much more important to follow a detailed design process for DC systems because of the potentially large cost implications of either over-dimensioned or under-performing insulation.
In the Cigré guidelines, designers are encouraged to follow an exhaustive approach with the aim of minimizing uncertainties in the input data and its impact on final design. The dimensioning principles are introduced on the basis of a flow chart (see Fig. 3) to provide a holistic overview and context for each of the activities that make up this process.
In this flow chart, the overall design strategy is shown in the vertical column of numbered blocks on the left. To the right of each block, a number of ways of obtaining the relevant information are presented. The further right one moves on the chart, the less certain are the results. For example, when determining site severity, information from existing DC lines will provide more accurate results than qualitative severity estimation. Briefly, the activities in the flowchart can be described as follows:
1. Identify Candidate Insulators
Due to the particular demands of HVDC, some insulator types are specifically optimized for DC applications with respect to the insulating materials and shed profiles used. Consequently there is only a limited choice of insulators available for DC applications. Despite this, choices still need to be made regarding the type, insulating material and profile shape for the insulators that will be utilized at a particular location. The initial selection of the candidate insulator is usually based on a simplified preliminary site assessment. Choice of candidate insulators could be revised throughout the process as more detailed information about site conditions and applications becomes available.
2. Assessing Environmental & System Stresses
Ideally, pollution deposition measurements over a period of a few years are needed to provide accurate data on the nature and severity of pollution at a given site. For DC applications it is important that these measurements be performed on insulators energized to a representative DC stress. Since this is not always possible, other techniques have to be employed to obtain this information.
If there is data on the performance of HVAC installations in the area, it may be possible to ‘translate’ this information to the ‘HVDC situation’. But this process is at best only approximate since the designer needs to make assumptions about the differences in accumulation on DC and AC energized insulators. It is also possible to use a general environmental assessment to identify a comparable environment in a different locality where an existing HVDC installation is in operation. Data from this installation could prove useful in design and selection of insulation for the new installation.
3. Determining Insulator Characteristics & Dimensions:
The most accurate way to select insulators for a new installation is to directly determine risk for flashover as given by service experience of DC lines and substations located in the same area or exposed to similar environmental conditions. This flashover risk can also be obtained through establishing test stations, where the performance of a range of pre-selected insulators is monitored under DC voltage at locations considered representative of the new line and station corridor.
Where there is previous experience with DC lines in the same area, excellent data on insulator performance will be available on which to base preliminary design. If there is a lead-time of one year or longer, useful data can be obtained from energized insulators installed in field stations located at representative sites along the length of the line and at the converter sites. Insulators at such field stations must be energized to representative stresses to take account of the influence of the electrostatic field on pollution accumulation, which can prove significant.
Instead of determining risk of flashover directly it is also possible to follow a simplified deterministic method for design. Here, the pollution stress, i.e. maximum pollution level on insulators, is determined from pollution measurements and through site condition studies. Insulator strength is then estimated from published information or based on performance data summarized in the guidelines, with several correction factors applied to correlate test conditions with site conditions. This data are then used to make a rough selection of insulator type, material and dimensions.
4. Design Verification
This is the last step in the process whereby the chosen insulation design is evaluated, either by comparison with past experience or through testing.
Simplified Dimensioning Process
Besides the detailed dimensioning methodology discussed above, a simplified method was also established. The basic intension was:
• To provide useful orientation at the start of a project to identify a range of possible preliminary solutions;
• To analyze outage performance and adequacy of the insulation solutions of existing systems.
It is important, however, to stress that the simplified method has serious limitations that could result in either over- or under-dimensioned insulation. It is therefore not considered accurate enough for the final design process. On the other hand, it does provide insight into the parameters that need to be considered when dimensioning insulators with respect to pollution for HVDC systems.
The first part deals with determining pollution severity of the site, which is the equivalent value of the ESDD at a reference NSDD value of 0.1 mg/cm2. The second step is to adjust this value individually for each type of insulator considered so that the required USCD can be determined. The two parts of this process are shown in Figs. 4 and 5 respectively.
Determining Site DC Severity
An overview of the process to determine DC severity is presented in Fig. 4. The aim of such an assessment is to obtain an accurate picture of the contamination severity of area concerned based on data collected over a relatively long period. The initial assessment is usually based on:
1. Collected performance data on existing lines or substations, preferably DC energized. AC data could also be useful;
2. Identification of the type (i.e. Type A or B, as defined in IEC 60815-1) and composition of pollution (i.e. type of salts, non-soluble deposits etc.);
3. Measurement of quantity of pollution present;
4. Characterizing climate, specifically if there is a prolonged dry season;
5. Assessment of geographical, topological and geological features to identify possible contamination sources; and
6. Survey of present and foreseeable future pollution sources and land use.
For critical installations (e.g. convertor stations), the above information might not be accurate enough, resulting in the need for more detailed assessment. This preferably includes setting up experimental stations at representative locations with a selection of DC energized insulators to obtain an estimate of long-term pollution accumulation.
As an alternative, it is also possible, but with increased uncertainty, to base the site severity assessment on measurements on AC energized or non-energized insulators. In such a case it is necessary to estimate the contribution of electrostatic field to accumulation on DC energized insulators. This is done with the DC/AC accumulation factor Kp (applicable to both ESDD and NSDD values). This factor can vary from 1 to 10 but, for a simplified dimensioning process, values between 1 and 3 are more typical.
Once the site severity measurements are available, the maximum value of the average ESDDs measured on the insulators is converted to an equivalent laboratory test severity. With this correction, it is recognized that artificial testing differs from natural pollution in a number of important aspects, namely:
7. Type of Salt: Laboratory testing is performed mostly with marine salt (NaCl) whereas natural pollution layers often contain less soluble salts such as gypsum (CaSO4). At present, however, there is no generally applicable method to quantify this effect other than performing specific flashover testing on insulators with natural pollution, so assume Kc=1.
8. Amount of non-soluble material present in pollution layer: The standardized laboratory test usually subjects the insulator to a pollution layer with a non-soluble deposit density (NSDD) of 0.1 mg/cm2. In service, NSDD levels typically vary from 0.01 to 10 mg/cm2. Measured ESDD is normalized to an NSDD value of 0.1 mg/cm2 with a factor, Kn.
These corrections result in an estimate of equivalent site severity at NSDD=0.1 mg/cm2, which is then defined as the DC site severity.
Determining Required Creepage Distance
Required insulator dimensions (notably creepage distance) are determined from available service experience or test results. If such data is not available, representative tests can be performed on candidate insulator types to determine their statistical flashover properties. By ‘representative test’ it is understood any laboratory test designed to imitate natural contamination conditions as closely as possible by replicating (1) pollution severity (i.e. ESDD and NSDD) (2) pollution composition, i.e. type of salt, and non-soluble components, (3) uniformity of the deposit and (4) wetting conditions.
Alternatively standard laboratory pollution test results can also be utilized but a number of adjustments to the DC severity then become necessary. These adjustments are dependent on insulator type and are therefore done separately for each type of insulator utilized (see Fig. 5).
Briefly, such adjustments can be described as follows:
9. Non-uniformity of the contamination layer: Flashover voltage of non-uniformly polluted insulators can be significantly higher than uniformly polluted ones. This effect is corrected for with a factor, Kcur.
10. Insulator diameter: Larger diameter insulators collect less contamination than small diameter insulators. This is corrected for with another factor, Kd.
11. Statistical considerations: This correction factor, Ks, is chosen to obtain sufficiently low flashover risk.
Application of these correction factors results in the design DC severity that corresponds to the pollution severity at which representative laboratory tests can be performed. At this stage, it is also possible to make a first estimate of leakage distance (i.e. unified specific creepage distance or USCD) required for the project, based on the graphs in Fig. 6. Two graphs are presented, one for hydrophobicity transfer materials (HTM) such as silicone rubber and one for non-HTM insulators such as glass or porcelain.
These graphs are valid for both line and substation insulators with relatively small diameter. Insulators with larger diameters generally have a lower flashover voltage than insulators with smaller diameters and therefore require longer leakage distances. This effect is corrected for with a factor, Cd. At present, it is not considered necessary to correct for the effect of diameter on hydrophobic insulators.
For installations at high altitude, an additional correction factor, Ca, can be considered to adjust the leakage distance for the lower flashover voltage under low air density conditions.
As confirmation of the general applicability of the simplified method, the Cigré WG analyzed actual outage performance of various installations against the design curves presented. In practice this proved to be rather difficult as detailed information on outage performance and corresponding pollution severities for existing HVDC systems are not generally published. However, some useful information was found and the results of this analysis are presented separately for substation (Fig. 7), line (Fig. 8), and HTM insulators (Fig. 9). In the graphs, a distinction is made as follows:
12. Design values: These data points represent minimum creepage distance values implemented on actual HVDC systems but for which performance data is not yet available.
13. Good service performance: Indicates minimum creepage distance values implemented on actual HVDC systems for which satisfactory service experience has been reported.
14. Flashovers: Indicates minimum creepage distance values implemented on actual HVDC systems for which unsatisfactory service experience has been reported.
The data in these figures confirm the appropriateness of the simplified method to obtain a realistic first estimate of insulation dimensions required at HVDC installations.
After years of work, Cigré has finalized a set of guidelines for selecting and dimensioning external insulation with respect to pollution. Of special interest is the latest Guide that describes the methodology and principles by which HVDC insulators should be selected and dimensioned for polluted conditions. Relevant technical background information – especially on parameters unique to DC – is provided in the Guide’s appendices.