Overall evaluation of the materials

Glass fibers with epoxy specific sizing from manufacturer 1 performed better than other glass fiber types in several test. The structures in which they were used have quite uni-form and even layers based on the microscopic examination, and they peruni-formed well in the dye penetration test. Flexural properties were among the best, but the compressive strength was worse to the other glass fibers, except the glass fibers from manufacturer 2. Glass fibers from manufacturer 2 were clearly inferior in other tests as well. Five-layer

structure showed most variation in dimensions and the layers of the epoxy-based tubes have major flaws. The performance in dye penetration test was also worse compared most of the other glass types with epoxy resin. Of mechanical properties, especially the compressive strength was lower. The structure with the only boron-containing glass from manufacturer 3 distinguished in some of the test. It had a uniform structure, but it was the only structure that could quite certainly be said to contain some seeds in the glass fibers based on the dye penetration test. It is however unlikely, that the seed level is high in case of boron-containing glasses since a high seed count is generally the problem only in boron-free glass fibers. From flexural properties boron-containing fibers were clearly worse than the boron-free fibers but the compressive strengths of the fibers did not differ. Structures with glass fibers from manufacturer 4 were average in many re-spects. It has quite uniform structure based on the microscopic examination, but dye penetration test showed some porosity. Structure has highest compression strength, and its flexural properties were between the structures with glass fibers from manufacturer 1 and 2.

In case of electrical insulation properties, the superiority of epoxy specific sizing was not seen. In the test before exposure to moisture, the leakage current of glass fibers from manufacturer 1 was lower than manufacturer 3 and 4 glasses. After moisture exposure, there were nearly any differences in leakage current between different glass types. The change in the leakage currents because of moisture exposure was greater for glass fi-bers from manufacturer 1 than glasses from manufacturer 3 and 4 but less than manu-facturer 2. Surprisingly in dry tested tubes the glasses from manumanu-facturer 2 have lowest leakage currents. Instead, the phase angle of five-layer structure was clearly lower and the increases in leakage currents after moisture exposure were higher than other glasses. The leakage currents of glass fibers from manufacturer 3 were lowest after the moisture exposure and the increase in them was smallest. However, it has the highest leakage currents and phase angles before exposure to humidity, which indicates that the electric current flows through the glass and it has poorer electrical insulation properties compared to other glasses. The same observations but at lesser extent also applies to the glass from the manufacturer 4.

It can be concluded that the sizing of the glass fibers influences many properties. The epoxy specific sizing seems to bring advantages in structural integrity and mechanical properties but its effect on insulation properties was not as clear. The sizing used by manufacturer 2 was clearly not as suitable for epoxies than the other multicompatible

sizings from other manufacturers, but it seems to behave well with vinyl ester and poly-ester. The chemical compositions of the glass fibers and their sizing vary significantly depending on the manufacturers and the formulations used are confidential information.

The most important is that the sizing is compatible with the resin, so that defects, which affects both the mechanical and electrical properties, are not caused in the laminates.

Another significant factor when considering materials for electrical insulators is the boron content of the fibers. The higher boron-content of glass fibers from manufacturer 3 was reflected in poorer mechanical properties. Removing boron and replacing it with titanium oxide has been shown to improve strength. Boron is also main factor exposing insulators to brittle fractures. It should be however noted that most of the literature research were related to unidirectional rods rather than hollow tubes where the fibers are wound with different angles and the tensile stresses of multidirectional tubes are not so large that the brittle fractures would be so likely. Studies related with boron content mostly focused on brittle facture or the effect of moisture, but not at the electrical insulating properties themself. No large differences in water absorption and leakage currents have been ob-served for boron-free and boron-containing fibers when the seed count have been low.

The high seed count instead has increased the leakage currents. The importance of the low seed count was emphasized in literature review and survey for the glass suppliers, but the experimental section of this work did not provide definitive observations on the effects of the seeds.

Of different polymer types, polyester is by far the worst option for electrical insulation.

Although it performed well in the dye penetration test and the structure appeared to be uniform under microscopic examination, its mechanical properties and electrical proper-ties were inferior to vinyl ester and epoxy. One of the most important properproper-ties of poly-mers in terms of electrical insulation is their moisture absorption and the importance of moisture exposure was especially emphasized in case of polyester-based tubes. When tested dry, the insulating properties of polyester-based tubes outperformed vinyl ester and most of the epoxy-based samples. However, polyesters have a high tendency for moisture absorption, which make it unsuitable for electrical insulation applications. Due to weaker chemical structure and the lower fiber volume fraction of polyester-based structures, its mechanical properties are also poor. Insulating properties of vinyl ester-based tubes were slightly poorer compared to epoxy-ester-based both before and after the exposure to humidity. Vinyl ester stood out in favour of its compression strength and structural integrity. Based on the literature review, vinyl ester has the best resistance to brittle fracture because of its fracture toughness and low amount of exposed fibers and interfacial splitting, but more research on brittle fractures of hollow tubes would be

needed for the results to be applicable to the studied products. The superiority of epoxy both electrically and mechanically proved obvious due to its durable chemical structure and moisture resistance properties. In addition, the advantage of epoxy is that its prop-erties can be altered by the choice of the base resin and hardeners. Careful selection of the hardeners is also important for brittle fracture resistance.

The number of layers of the structure is not very meaningful in terms of insulating prop-erties. Layer number did not correlate with leakage currents before exposure to moisture and after moisture exposure the leakage currents of three-layer structures did not differ from five-layer structures. Lower number of layers impaired the flexural properties, but in terms of compressive strength they proved to be the best. Instead of the number of lay-ers, structural integrity of the structures is more important. The addition of glass tissue lowered the leakage currents in dry tested tubes, but the increase in leakage currents after moisture exposure were largest of all, making them slightly less suitable as electri-cal insulators than the structures without glass tissue. From the point of view of mechan-ical properties, the addition of glass tissue to the structure would need further investiga-tion, because in most of the test, one structure with added glass tissue behaved differ-ently than the other, so no conclusions about the effect of glass tissue on mechanical properties could be made.