The results of natural ageing tests have shown that other design parameters than the creepage distance may be critical for the short and long term performance of composite insulators
such as: the design of the triple junction, the form and direction of the moulding line and the distance between the bottom flange and the first shed. Based on finite elements method field calculation, the paper illustrates how the insulator design, in general, and above mentioned specific design items, in particular, may influence the behaviour of composite insulators, providing confirmation and explanation of the results of natural observations. The approach illustrated in the paper may assist in extending and improving the concepts and criteria used for an effective design of composite insulators.
Polymer materials such as silicone elastomers,hydrocarbon elastomers and epoxy resins are being increasingly used instead of porcelain and glass for outdoor insulation applications such as line insulators, bushings
, hollow core insulators, cable terminations, etc., mostly due to the following advantages:
light weight = lower construction and transportation costs;
vandalism resistance = less gunshot damage;
high strength to weight ratio = longer spans/new tower design;
improved transmission line aesthetics;
unexplosive housing = improved safety for the staff in the station and for the installation equipment.
and glass these polymers have low surface free energy which makes the virgin surface (new and without exposure to the environment) of the polymers inherently hydrophobic (water repellent). Hydrophobic surfaces present a higher resistance to leakage current flow than porcelain or glass surfaces (hydrophilic surfaces). They also require higher current and commensurate energy dissipation to initiate the well-known phenomenon of dry band arcing which, ultimately, is responsible for material degradation in the form of tracking and erosion. The lower leakage current and consecutive lower probability of dry band formation require a higher applied voltage to cause flashover. In one word, due to the hydrophobic surface, the polymer materials typically offer much better contamination performance than porcelain and glass . Service stresses, such as surface discharge activity on hydrophobic surface, UV exposure and chemical attack, cause the reduction or complete loss of hydrophobicity and dry band formation under the same process as porcelain or glass. It has been observed that, in case of silicone rubber, the surface, due to the diffusion of mobile low molecular polymer chains (LMW) from the bulk to the surface  and the rotation of surface hydrophilic groups away from the surface , recovers hydrophobicity when there is little or no dry band arcing .
The ability of the material to control leakage current, which represent the first defence line of the insulating device, varies significantly based on polymer material used, but also on its interaction with the product design. However, even housing materials that have a tendency to recover their lost hydrophobicity must be able to withstand dry band arcing without tracking or erosion a secondary line of defence against contamination induced flashover. Housing design can also influence leakage current during the periods of reduced or lost
hydrophobicity. Therefore the key to longevity in polymeric (non-ceramic or composite) insulators is to ensure that leakage current is kept low. Housing material formulation and leakage current path design are two interdependent tools that manufacturers have available to solve the performance optimization problem. Moreover, design weaknesses (lack of voltage stress relief, poor sealing between materials and connecting hardware, improper method of coupling the endfittings
) as well as the quality control problems play a very important and probably a primary role, in determining the life time of these insulators. As we just said, polymer materials usually outperform porcelain and glass in contaminated environments, but they must be adequately designed and manufactured to withstand such conditions without accelerated ageing (in dry and non–contaminated environments these insulators normally have a very long life).
To summarise our earlier discussion which shows that housing polymer formulation, product design, and manufacturing process are interdependent and that manufacturer, in order to offer a good insulator, has to solve the higher order optimisation equation, we shall use our extension  of a matrix developed by Prof. H.Kärner  .
As we can see, it is quite normal to produce a bad insulator from bad materials as well as from a good material with a bad design. To produce a good insulator from a bad material, even with a good design, is virtually impossible. However, even if one starts with a good material, there is still a possibility of manufacturing a bad insulator if the design is poor. Finally, one can produce a bad insulator starting with a good material and having a good design if there are poor manufacturing process and/or poor quality control. Obviously, a good composite insulator could be obtained with a perfect combination of design and polymer material formulation. To transfer the "could" to the "can", the third condition has to be fulfilled manufacturers know–how. This paper discusses some design aspects, such as the moulding line and the optimal distance between the first shed and the metal flange, including the triple junction point.
Fig.1 shows three principal designs of composite insulators. Insulators according to Fig.1a consist of a fibre reinforced polymer (FRP) rod (tube in case of hollow insulators) covered with a seamless sheath. An extrusion process used in manufacturing of cables applies the sheath. For the reason of bonding, a primer is applied to the rod surface prior to extrusion, enabling the sheath to obtain chemical cross-linking to the rod surface. The sheds are moulded separately and pushed onto the sheath by means of a slippery vulcanizing paste. When the requested number of sheds is positioned as designed, sheds and sheath are vulcanized together at elevated temperatures (HVT – high temperature vulcanisation). The bonding between fittings and housing is realised using metastable silicone rubber sealing. Insulators acc. to Fig.1b are produced in a single shot moulding process where FRP rod is positioned between two halves of a parted mould and the housing (including the sheds at the same time) material is injected into the mould. Due to heating, the vulcanising process starts to rosslink the housing materials as well as the bond between housing and the rod surface. When a stable state of housing material is reached, the mould is opened and insulating body is taken out. The design of composite insulators acc. to Fig.1c uses modular weathershed housings including a number of the weathershed in a single module. The modules are then mechanically bonded to the adjacent module by an external polymer collar. The modules are mechanically sealed to the end fittings within an integral grading disk. The modules are assembled to the rod using a high dielectric strength silicone compound in the interface. The silicone compound is held in place by internal orings moulded into weathershed housing. All three designs are strongly related to the manufacturing process, each having their specific technical and economic advantages and/or disadvantages.