External insulation is one of the critical design aspects for HVDC equipment at converter stations. This is especially the case for UHVDC systems. The basic design principal for such a system with energy transmission capability up to 10 GW is to achieve superior reliability. This INMR article, contributed by Dong Wu of ABB HVDC, offers an overview of these issues based on R&D findings as well as design and operating experience of existing UHVDC systems. The issues addressed cover ambient conditions, type of insulators, dimensioning of air clearances, corona and electric field at converter stations.
External insulation is a critical design aspect for HVDC equipment at converter stations and especially so for UHVDC systems. Although voltage level is a main concern, with increasing voltage and size of equipment, other constraints such as mechanical strength and thermal effects also become more significant in design. Some, such as the effect of multiple gaps, become critical for UHVDC. Moreover, even design practice for personnel safety used at lower voltage levels may need to be re-considered. External insulation is a wide area that includes insulation design for air clearance, corona shielding and insulators for both indoor and outdoor atmospheric conditions. This report provides a general review of these and related issues.
Critical Issues re External Insulation
a) Site Conditions
Site pollution severity of a planned converter station could be estimated by different methods. Since there are already many HVDC systems in operation under different site conditions, one effective way is to study how pollution has affected these systems. Another is to set-up a test station near the location of the planed converter station. For UHVDC, it is even more important to get an accurate estimation of site pollution severity since decisions made based on this can significantly impact reliability as well as cost of a project. Other conditions that impact insulation design are wind speed and seismic requirements as these parameters will determine the mechanical requirements for equipment and become a constraint in insulation design.
b) Indoor Conditions
At converter stations, some equipment, such as converter valves, are generally located indoors. Other equipment is often located outdoors but can also be installed indoors i.e., inside a so-called indoor DC yard. Although the indoor condition is less complicated than that of outdoors, it is still influenced by the ambient air with different temperature, humidity and different levels of dust accumulation. These ambient parameters therefore need to be defined and controlled.
c) Selection Between Indoor & Outdoor DC Yards
For HVDC systems, an indoor DC yard is sometimes built for extreme site pollution severity or for stringent requirements of compact design. For UHVDC systems, the length of equipment has arrived at a value where mechanical integrity is critical. Considering the large amount of energy transmitted, an indoor solution may therefore become attractive. Still, such a decision requires a thorough study and comparison of different alternatives.
Type of Insulators
a) Surface Materials
For simplicity and mainly from a pollution performance point of view, IEC has divided the surface characteristics of insulators into HTM (hydrophobic property transferable material) and non-HTM. HTM includes mainly insulators with a silicone rubber surface. Non-HTM includes those insulators with surfaces that cannot maintain or recover their hydrophobic properties after being covered with pollution. Based on today’s technical developments, almost all insulators used in UHVDC systems have an HTM surface. For information related to the advantages and disadvantages of silicone rubber insulators made using different types of materials or with different types of manufacturing techniques, this is often too generalized and could be misleading. Any such discussions should therefore be based on a comprehensive study of operating experience. It is also inaccurate to generalize the hydrophobic property of an insulator by using only HC (hydrophobicity class) levels. HC measurement is a simple and quick method to get a rough estimate of the status of a surface. However, reports from on-site HC measurements have shown that HC levels are in most cases different at different points along an insulator and also can change quickly with ambient conditions. Expected pollution performance of an insulator cannot be obtained based on just a few HC measurements.
In the case of 800 kV DC, the relationship between dielectric strength and length of a station post insulator under artificial pollution had been confirmed to be linear when the applied pollution was higher than 0.05 mg/cm2 of SDD. But for lower SDD levels, this same relationship was slightly non-linear. This was obtained under conditions where the test object was a porcelain insulator with uniform pollution layer. For voltages higher than 800 kV DC, performing such tests becomes difficult. One of the difficulties is that fog applied in a test chamber tends to be non-uniform, which will lead to a partial wetting of the insulator. UHVDC systems have to and will be designed with the assumption that the linear relationship remains valid, even at higher voltage levels.
c) Shed profile
The importance of shed profile to insulator performance under various types of wetting conditions has been established, although it has been difficult to represent natural pollution conditions during such laboratory tests. For DC applications, in order to avoid bridging of sheds by partial arcing, relatively large shed spacing, compared to what is used under AC, is necessary. This is especially true for UHVDC systems, with their very large insulators. This principle, established in the case of porcelain insulators, also has to be applied to insulators with an HTM surface.
Based on the same profile principles, use of spiral or helix formation of sheds is related only to manufacturing techniques. For station insulators with large diameter and for insulators with an HTM surface, it is evident that, given the same profile parameters, sheds formed by a helical technique have given satisfactory performance.
d) Creepage Distance
Although it had been successfully tried earlier, it was in China that researchers as well as power utilities adopted, in a large scale, the principle of using shorter creepage for HTM insulators than normally used for porcelain insulators. Operating experience has been good. Indeed, this principle often made it possible to produce equipment for UHVDC applications. This principle can therefore be applied to most inland conditions. For extreme conditions with frequent appearance of wet pollution, e.g., close to the sea, restraint may need to be exercised given the risk of premature ageing.
For indoor conditions marked by absence of the most common wetting processes, such as rain, fog and snow, much shorter creepage than that for outdoors can and has been used, even though dust may appear indoors. In most of cases, creepage will not be the dimensioning parameter for indoor conditions.
For insulators installed in a horizontal position, such as wall bushings, shorter creepage than that used for vertical insulators can be used because of more effective natural washing and less risk of water bridging the sheds. This finding that shorter creepage can be used is especially true for wall bushings with silicone rubber sheds. The HTM surface makes a wall bushing insensitive to uneven rain. Operating experience of wall bushings with silicone rubber sheds supports this conclusion.
For apparatus with controlled internal voltage distribution, the dynamic change of external voltage distribution due to different weather conditions needs to be well considered in design. The tendency of using very long creepage could actually worsen the situation.
The main constraint for using longer creepage distance in a UHVDC system is the requirement on mechanical design. From a system reliability point of view, mechanical safety carries more weight as a risk factor than flashover of external insulation. A good balance must therefore be achieved between requirements on external insulation and the negative effect on mechanical strength of having too long insulators.
To better understand the relationship between stresses and strengths of external insulation, numerous laboratory tests have been carried out. These have provided valuable knowledge for design of external insulation. On the other hand, laboratory test cannot fully represent natural operating conditions, such as the complicated condition of pollution performance. Therefore, to rely on artificial pollution tests to determine the ‘go’ or ‘no go’ for an external insulation design may prove highly inaccurate. Accuracy will be even worse when HTM insulators are being judged by such tests. As such, it should be concluded that testing where R&D is the aim is beneficial for design while testing, (e.g. in form of type tests) to verify a design may often be misleading. For UHVDC, even availability of a facility to test full-scale equipment becomes a challenge.
Determination of Air Clearances
a) Multiple Gaps
Multiple gaps refers to the situation where one high voltage electrode faces several grounded electrodes, e.g., walls. This is not the same as for parallel gaps which refers to a situation that several high voltage electrodes stressed under the same voltage each face a ground electrode and act independently. For parallel gaps, there is a well-established way to account for changes in breakdown statistics in comparison to a single gap. For multiple gaps, however, this has to be handled by a well-coordinated insulation design. For UHVDC, with long air clearances in the main gap, it becomes difficult to exclude the influence of other grounded electrodes to the main gap.
b) Saturation and damaged electrode
Switching impulse is, in most cases, the dimensioning stress for air clearances at HVDC stations. It has been established that, under switching impulse, relationship between dielectric strength and gap distance is non-linear and also that it saturates rather quickly. This is especially the case for ‘rod-plane like’ gaps with large electrodes. For such gaps with large electrodes, these saturate even faster than the case for rod-plane gaps. Therefore, any improvement achieved by an increase of the radius of the electrode reduces quickly with increase in gap distance. This makes it necessary to improve the electrode each time voltage level increases. Otherwise, clearance has to be determined as for a rod electrode.
Moreover, even slight ‘damage’ appearing on the surface of the large electrode will cause drastic reduction in dielectric strength. For outdoor conditions, the combination of rain with pollution drifting from the surface of the electrode will make most of the air gaps behave as with a rod electrode. For indoor conditions, a ‘perfect’ electrode may also not be a correct assumption. In this case, installation structures such heads of screws will also ‘damage’ the electrode. Insects can appear indoors, which may also cause unexpected breakdowns. All these factors need to be included in insulation design.
c) Safety distance
The other effect of the multiple gap situation is that the breakdown trajectory may extend itself to reach other grounded electrode at a much longer distance from the main gap. This makes the ‘safe area’ disappear in the vicinity of the main gap. Traditionally, for a lower voltage levels corresponding to 500 kV AC system or lower, it was believed that with a metallic pedestal of some 2.5 meters, it is safe for station personnel to access the substation even during operation. However, for UHVDC systems, the arcing distances of insulators will be in the range of 8 meters and higher. The probability of breakdown taking place between the top electrodes to the ground will increase and working during operation under a pedestal of 2.5 meter can therefore not be considered safe for personnel.
It should be considered that any safe distance must have zero breakdown probability, i.e., with a distance that is evaluated with a margin of 5 sigmas from the U50. This criterion should apply only if, in extreme cases, a person must come into a DC yard during operation. The other measure would be to build a safe corridor, shielded from the arc of an eventual breakdown. Further study on safety measures becomes an absolute necessity.
For development of products for new and higher voltage levels, laboratory experiments are a necessary tool. Insulation design is based on the principle of probability. The probability of failure can be investigated with laboratory experiments using test procedures, e.g. the up-and-down procedure. The up-and-down procedure is an effective and relatively accurate procedure to determine withstand or breakdown probability. But various withstand procedures with small numbers of breakdowns cannot give the same level of accuracy unless a great number of voltage applications are made.
On the other hand, it is clear that equipment made of both self-recoverable and non-recoverable insulation cannot withstand many breakdowns. Some balance must therefore be reached that allows a reliable product to be developed while seeing not too many test objects being destroyed. Although compromise is necessary, it is not sufficient to justify any new design only by performing a standard withstand type test. Computer modeling in this case is a useful alternative. Combined with laboratory experiments, computer modeling of the breakdown process provides insight into this complex phenomenon. From an engineering point of view, although modeling techniques developed today may still not be so handy as a design tool, it is still valuable for simulating various situations that may not be well simulated in a test laboratory.
For testing on self-recoverable insulation, there have been challenges when testing at UHVDC voltage levels. It has been observed, for example, that discharges or even breakdowns could appear on other parts of the test circuit instead of the test object. Such discharges and breakdowns should be avoided by modifying the corona shielding of the test circuit. To facilitate observation of discharges, digital cameras with a photo acquisition system have been used. For each voltage application, photos over the test circuit from two different angles have been taken simultaneously with voltage application and immediately available for inspection in the control room. This makes it much easier for researchers to decide how to proceed.
DC Corona & Field
For indoor equipment and bus works, design should aim to be corona free. Although dust in the air, charged by DC voltage, may accumulate on the surface of electrodes, such effects should have been taken into consideration in design. For a corona free design, electric field will only be geometric and can be calculated without considering ion flow. For outdoor conditions, equipment and bus work are conventionally designed to be corona free in fine weather and when new. It is unrealistic to design for corona free for outdoor conditions. Ion currents appear when corona discharges are present. However, a converter station with high voltage equipment and bus work is not a public area and no people should to be present under ordinary conditions. Therefore, the criteria for ion-fields under DC transmission line need not be applied here.
Superior reliability for UHVDC systems can only be achieved by:
1. Utilizing the knowledge and operating experience from existing HVDC systems;
2. Identifying and meeting new challenges linked with increased voltage levels;
3. Balancing the requirements for insulation and other constraints in design.