4.1 Operation environment
Materials and structures of the insulators should have a lifetime of 20 to 30 years in challenging environments and extreme temperatures [6]. Insulators are subjected to me-chanical, electrical and environmental stresses that can lead to mechanical and electrical failures [5].
The environment in which the insulator operates have a great impact. In contaminated environments, insulators performance can deteriorate. [2] In operation various factors, like static and dynamic loading, electrical discharge, electrical current, UV radiation, hu-midity, rain and pollution, etc. affects to insulators performance. [31]
4.2 Failures
In most catastrophic cases, the insulator carrying power line breaks mechanically and failure lead to the drop of the energized transmission lines. In addition to the discon-nected line, freely hanging energized conductors cause a serious safety risk to people.
Another undesirable situation is an electrical failure that causes disconnections in the network. [4]
Most of the insulator failures are caused by bad design, material selection or quality control. In this regard, most manufacturers today produce high quality products, and the future failures depend primarily on the aging of materials. Degradation process of com-posite core is not common, but it may result in mechanical failures. End-fittings may slip out of place because of degradation. In addition, GRP core can break normally due to overload, or a brittle fracture can occur in the core. The lifetime of the insulator usually remains unchanged below the damage limit, where the fiberglass begins to break. How-ever, brittle fracture can be caused also at low loads. [4]
4.2.1 Brittle fracture
The most important failure mode of composite insulators is electro-mechanical brittle fracture (BF) failure, where stress corrosion cracks are formed inside the rod [5]. If the composites are subjected to pure tensile stresses parallel to the fibers without any acids, the fracture surface is broom-like. It consists of multiple fractures, fiber pull-outs and matrix cracking. In the BF process, where in addition to the tensile stress, the composites
are exposed to acids, the fracture surfaces are planar and run perpendicular to the fibers.
The number of fiber pull-outs and debondings is very limited and fiber and matrix are almost in the same plane. [8] The size of the smooth planar region can reach even 80%
of the cross-section of the rod [7]. Failure occurs typically either just outside the ener-gized fitting or inside the fitting. The morphologies of the fractures vary according to their location. If the failure occurs outside the fitting, it is typically characterized by severe resin decomposition. If the failure is inside the fitting, no significant decomposition of the resin usually occurs. Severe surface contamination is typical and can be so strong that the single fibers on the fracture surface cannot be detected. [8]
In microscopic level, the formation of mirror, mist and hackle zones on the fracture sur-faces of single glass fibers is typical [8]. Mirror zone is smooth and circular, and it typically contains the starting point of the fracture of the single fiber. There the crack propagation is relatively slow. Mist zone has matt and rougher surface than mirror zone. It is a tran-sition region between mirror and hackle zone. There the propagation of the crack be-comes rapid. Hackle zone is rough-textured, and it has radical irregular fracture lines running from the mirror zone, indicating rapid crack propagation. Hackle lines run in the direction of crack propagation. [32]
4.2.2 Cause of brittle fracture
The brittle fracture of composite rods has been widely accepted to be caused by stress corrosion cracking (SCC). SCC can occur when the rod is simultaneously exposed to tensile stress and corrosive environment containing free hydrogen ions. [8] At the mo-lecular level, mechanism of SCC is described by ion exchange reaction. Non-siliceous oxides, like the large metal ions of glass fibers (Al, Ca, Fe, Mg) are leached out and replaced by hydrogen ions. It is considered that leaching of calcium and aluminum is the main reason for acid corrosion, because their amount is about 35 wt% of the glass, while all non-siliceous oxides together make up 45 wt%. [33] Leaching weakens the fibers and fractures can form if low tensile stress is applied along the fibers [8]. BF can form at loads that are only 5 to 20% of the maximum load-bearing capacity of the composite rod [7].
It is generally approved that BF is caused by the simultaneous application of a tensile stress and an acid. However, the source of the acid controversial. Theory of Montesinos et al. [34] claimed that SCC can be caused by hydrogen ions in terms of water and that BF can happen due to water and mechanical stresses, and it occurs more probable with water than acids. [35,36] However, that theory has been questioned by the study where water could not cause SCC without either high electric fields or heavy contamination [36].
According to Tourreil et al. [37] the source of the acid is either nitric acid formed in service or carboxylic acid generated by hydrolysis.
Based on the nitric acid theory, corona or dry band discharges in presence of moisture and oxygen can produce nitric acid. If the housing is damaged and the core is exposed to the atmosphere in area where electric discharges may occur, nitric acid can be formed.
[37] Nitrogen oxides are formed as a by-product of gas discharges and when they react with water nitric acid is formed [8]. This theory is supported by the observations that most of the BF failures have appeared just outside the live end-fitting and some inside the fitting. Failures have not occurred close the grounded end or in the distance from the fittings. This scenario also explains how the higher number of BF failures is related to higher transmission line voltages. [38] In many cases the housing is damaged, and rod exposed. This favours the theory that acid comes from external sources rather than in-ternal ones. [39,40] The identification of nitrate from the surfaces of failed insulators have confirmed the generation of nitric acid in service. The process has also been simulated under laboratory conditions. [38]
According to carboxylic acid theory, water can under certain conditions, affect all epoxy, vinyl ester and polyester resins and form carboxylic acid because of hydrolysis [37].
Epoxy formulations consist of epoxy and hardener. Epoxy formulations are discussed more in chapter 3.2.3 Epoxy. Most common hardener belong to phthalic anhydride group. Hydrolysis of the hardener transfers them into acids. Acids can be generated be-fore or during the fiber impregnation and picked up by the fibers when they pass through the resin bath in the pultrusion process. In the finished product, clumps of hydrolyzed hardener may be randomly located on or near the surface, or somewhere inside the core.
If water meet acidic clump, it enables the acid to enter the cracks and start the BF pro-cess. [39,40] In the other case small amount of unreacted hardener remains in the fin-ished pultruded rod. This can happen if the resin is not completely cured due to too short curing time or a low temperature. [37] If water reach the rod during service life, it will hydrolyze the hardener still present. This newly formed acid can, in combination of tensile stress, lead to BF. Water acts as a carrier for the acid and without water in contact with the rod, the occurring of BF is very unlikely even in the presence of acid clumps. [40]
It has been confirmed by several studies [37,39,40] that hydrolysis of the hardener can turn them to acids. The reaction is described in Figure 9. Type of the hardeners used in the studies is not mentioned but most likely they belong to anhydride family, which is mentioned in the articles as the most common one. Laboratory test proved that clumps of hydrolyzed hardener can lead to BF in the presence of moisture [39]. Field case study
shows that same process explains field failures. Fourier transform infrared spectroscopy (FTIR) analysis of failed insulator shown that small amount of unreacted hardener was present in the rod and presence of the acid that is obtained by the hydrolysis of hardener.
[40] Field failed insulators, including polyester, vinyl ester and epoxy-based materials, were studied with FTIR analysis to detect presence of nitric and carboxylic acid. Of 18 field failed insulators, 11 failed because of carboxylic acid, 2 because of nitric acid and 5 cases were not clear. Based on the study, it is 5-10 times more likely that BF is caused by an acid that is a by-product of the manufacture of the rod itself, instead of nitric acid occurred in service. [37]
Figure 9. Reaction where hydrolysis of the hardener leads to creation of carboxylic acid [40].
Observations that support the carboxylic acid scenario rather than nitric acid scenario are that BF have occurred even though the housing or the fitting seal that prevent water from entering the core have not been damaged, and BF have occurred far away from energized end fitting where electric field is very low, and expose rods have not fail even they have been years in service exposed to environment and acids. Based on that the source of acid cannot logically came outside the insulator but must be internal. [39,40]
The theory is also criticised. According to Kumosa et al. [38] it is chemically possible only with epoxy/anhydride resin systems and most of the failed insulators have been E-glass/polyester rods and failures of E-glass/epoxy rods is rare. Nor can the model explain why the failures typically occur outside the fitting. Failures should be able to appear at anywhere along the core, but the location typically correlates with electric field concen-tration.
4.2.3 Prevention of brittle fracture
Common things in the models are that BF is caused by SCC and that water ingress initiates the process [35,36]. BF can be prevented by not allowing water or acid inside the insulator or designing the rod with very high resistance to SCC. Even the best pro-tection against moisture fails over time. It would be more efficient to design the composite core to be SCC and BF resistant. [35]
Among glass fibers, resin and their interface, the glass fibers have been found to be most vulnerable to SCC. [33] Fiber and resin type, resin fracture toughness, surface fiber ex-posure, interfacial strength and moisture absorption affects BF resistance. [41,42]