Residual Voltage Endurance of Generator Insulation Systems

Conference Proceedings of ISEIM 2014

Residual Voltage Endurance of Generator Insulation Systems Christof Sumereder1 and Mario Dolcic 2 1

Graz University of Technology, Institute of High Voltage Engineering and System Management 2 Salzburg AG *E-mail: [email protected]

Abstract—The insulation system of a generator is stressed by electrical, mechanical and thermal loads during the whole lifetime. By the means of diagnostic measurements the worsening of the condition within the dielectric system can be evaluated. To ensure the operational safety for the next maintenance period a hipot test with raised level can be applied. In spite of all carefulness it can happen that a breakdown of the insulation system happens. Beside the costs and required time for machine repair contractual damages can result in high costs for the operator. The idea of these investigations was to examine the residual voltage endurance of aged generators. For this reason operational aged generator bars were dismantled carefully from two stators of same type which were set out of operation for renewal. The generator bars were tested by the means of following dielectric methods: Insulation Resistance, Dielectric Dissipation Factor and Partial Discharge as well as accompanying material tests. The test objects were electrically aged with a three time raised nominal voltage with an expected lifetime between 100 and 200 hours. In regular intervals the dielectric tests were repeated periodically. In this paper elected results were presented with the focus on the diagnostic expressiveness of the investigated dielectric parameters. Statistic methods were used to describe and display the aging behavior of the investigated insulation system.

Keywords: voltage endurance test, lifetime, generator bar, dielectric diagnosis, condition evaluation

I.

INTRODUCTION

In these investigations the residual lifetime of operational aged generator bars and the dielectric behavior of the insulation systems during the aging process were observed. In the beginning the state of the art in aging mechanism of mica insulation systems was the focus of the scientific preliminary enquiry. In [1-4] typical phenomena of failure types in resin mica insulation systems were discussed as well aging mechanism and procedures and diagnostic methods for condition evaluation were presented.

Figure 1. Cross section of investigated generator bar and microscopic view of the structure

18.000/15.000 starts. The insulation system was identically: micatherm® without inner corona protection, insulation class B, Resin Rich technology. The technical specifications were: rated voltage 10,5kV, rated power 32MVA both generators were operated in different storage power stations of the same utility. At the generators diagnostic measurements were applied in regular intervals of approximately five years depending on the hours in service and number of starts. Following measurements were applied as routine tests: insulation resistance (IR), dielectric dissipation factor (DDF) and partial discharge (PD). The results can be used for the interpretation of the condition with the special focus on the trending of each parameter. In the following diagrams the diagnostic parameters should be observed within the last five years of machine operation. At first the insulation resistance is investigated in below diagram. No signs of pollution of surface contamination can be derivate from the test results. The insulation resistance is constant over the whole diagnostic interval.

For the machine operator the main questions were how long the aged machines can be used, which risk of outage is given and which diagnostic methods were most reliable. To answer these questions in consideration of electrical aging these tests were carried out on operational aged generator bars. II.

TEST OBJECTS

The generator bars for the investigations were dismantled at two hydro generators of the same construction type. When the machines were put out of operation they showed an age of 44/43 years with 116.000/105.000 hours in operation and

Figure 2.

Insulation Resistance within the last 5 years of operation


Figure 3. dielectric dissipation factor during machine operation

As next point the historic results of the dielectric dissipation factor measurements were shown in Fig. 3. It can be observed that the DDF is rising with increasing operation hours and starts by trend. The last dielectric inspection was done immediately before dismantling of the stator bars (green line) in 2011. The DDF is strong dependent from generator temperature, so it is important to protocol the present temperature to avoid misinterpretation. One possibility to beware of the temperature dependence of DDF is to look at the so called Tip-Up. In Fig. 4 the historic Tip-Up values were shown in a diagram. It can be clearly seen that the best values were at lowest operation hours (orange line) and the trend goes to higher values with increasing hours in service and machine starts (green line). To evaluate the results concerning the condition the absolute level according to the standard EN 50290 or recommendation can be done or a relative observation by the change of DDF by time is possible. For the interpretation of the electric aging tests the relative method was chosen because of the better explanatory power.

Figure 5. PD within the last years of machine operation

III.

ELECTRICAL AGING OF GENERATOR BARS

From earlier investigations [5] the height of the test voltage respectively the level for the voltage endurance test could be determined by 3 times rated voltage for an expected residual lifetime between 100 and 200h. The generator bars were assembled in a slot simulation. The test setup is shown in Fig. 6. Each 40h all diagnostic measurements were executed. Over the whole aging period the temperature was constant. IV.

TEST RESULTS

At first the Insulation Resistance (IR) and the Polarization Index (PI) should be observed. The results during the aging process can be seen in Fig. 7. In this diagram the IR after 10 minutes were displayed (red line). It can be seen that the IR slightly decreases from 2.000GOhm down to 1.000GOhm. The PI showed a minimum of 3,5 after 40h and a maximum of 4,5 after 77h.

The partial discharges also were measured during the revisions of the machine. In the following diagram the development within the last five years of machine operation is given. The absolute level of the apparent charge is qualitative compared to other machines of same type. From the qualitative behavior of the pd over the test voltage a partial discharge inception voltage (PDIV) at 40% of the rated voltage and a strong raise of pd from 80% of rated voltage can be observed.

Figure 4. dielectric dissipation factor Tip-Up during machine operation

Figure 6.

test setup for electrical aging


Figure 7. IR and PI over aging period

The influence of temperature can be neglected because it was documented in a range of 24°C ± 2°C, so that there should be no necessity for a thermic correction. The measurements were done according to the IEEE 43-2000. The measured current consists of four components: leakage, conductance, capacitance and absorption. The absorption is dominated by orientation polarization and displacement. A healthy insulation system shows a characteristic low leakage, very fast decreasing capacitance and very low conductance behavior. The absorption is strong dependent to the used material (former insulation systems asphalt, today resins) as well as the condition (moisture, clean surface, surface tracking). The surface current is at a clean and dry system stable after approximately 4 minutes, if this time is significant shorter (