Fiber Bragg grating monitors for thermal and stress of the composite insulators in transmission lines

1 Introduction

With many advantages of resistance to flashover incidents in transmission lines, composite insulators have been widely utilized, especially in sand hazard areas. With the implementation of the global energy interconnection,composite insulators will be much more necessary for the engineers. Nevertheless, due to mechanical, environmental and manufacturing effects, composite insulators are prone to damage and failure, due to brittleness of the insulator rods, seepage of moisture from end-caps, as well as accumulation of salt, dust and ice on the composite insulator elements.

The running status of composite insulators has attracted much attention of engineers and researchers to the failures.The investigators have developed many methods to get the rod fractures, the end-fitting slippages, the mechanical strength declines, contamination levels, and so on. Many stress tests were executed by Liang, in which experimental rods with different materials are used for the insulators to get the environmental effects including temperatures,chemical corrosions, electric fields, and gale actions [1].Ahmad developed artificial neural network (ANN) to accurately model the influence factors on equivalent salt deposit density (ESDD) to estimate critical pollution levels of ESDD, which provided a competent method preventing flashover of insulators [2]. Optical fiber technology was used by Lepley to inspect and pinpoint the fault point of composite insulators. Optical fiber sensors were spirally installed in the rod [3]. An integrated optical sensor was designed by Rong to detect electric field distributions of the insulators with the running status [4]. Quartz fiber sensor with low loss was favored by Cai to obtain ESDD of insulator, and this method was based on the relationship between light energy loss caused by ESDD and optical field distribution in the dielectric waveguide [5].

The fiber Bragg grating (FBG) is a sensor with high sensitivity to temperature and stress changes, which is used to measure thermal and strains nearby and obtain the physical and chemical information related to temperature and stress indirectly. With the advantages of tiny size, excellent insulation and resistance to electromagnetic interference, this sensor has been extensively popular in aviation [6], energy[7] and construction [8-10]. In the power delivery areas, the application of FBG is in initiative stages [11-13]. The FBG sensor attached at the skin of insulators is developed by Ma for monitoring ESDD with addressing temperature changes caused by the insulator pollution flashover [11], but this design changes the insulation performance of insulators.With the application in composite insulators, the FBG sensor is embedded in hollow structure to support the composite insulator and monitors the temperature changes caused by pollution flashovers [12, 13]. Based on works of [14], The aims of this paper are as follows:

(1) Mechanical test on running composite insulators

500 kV insulator samples from a polluted area provide defect points of composite insulators, and these points are the monitoring keys.

(2) Relationship calculations

Thermal and stress are two fundamental monitoring parameters of the composite insulators. Calculations are done to get the effects of the thermal/strain changes on the center wavelength of FBG embedded in insulators.

(3) Feasibility design and calibration tests

Based on fault experiment results and calculations,this paper introduces the development of the composite insulator with FBG implanted, including the design of optical sensor system, the molding process of the composite insulator with FBG implanted and the design of online detecting system. Calibration test is done in this work,which includes the temperature and stress test.

(4) Application tests

Application tests of this monitor system are performed in the field of transmission lines. The system has detected thermal and stress changes of composite insulators caused by a strong gale.

2 Typical defects in composite insulators from polluted areas

For the mix of chemical, mechanical, and electrical factors, the typical defects of composite insulators include the sealing of the end-fittings and mechanical strength of the rod materials [1]. In order to get the typical defects,500 kV insulator samples are randomly collected from a polluted area by mechanical tests, and the defects are the key monitoring points.

2.1 Sealing performances of the end-fittings

According to the experimental method from IEC 61109-2008 [15], sealing performances of the end-fittings are analyzed as:

(1) The end-fittings of the insulators are cleaned up as the samples;

(2) The magenta with the concentration of 1% is uniformly painted as the mark at the surface of the endfitting edge;

(3) After 20 min, automatic extension testing system(RWL-4500, Reger, China) is applied to test the sample with 70% (147 kN) of the rated mechanical load (210 kN);

(4) After 1 min, the sealing performances of the endfittings are checked by the painted magenta.

The results of the sealing performances are shown in Table 1. The experimental data show that the possibility of the sealing defects in the end-fittings has positive relationship with the running time of insulators, and the defect possibility is about 100% with the running time above 8 a; the defect possibility of the socket (high voltage part, which is shown in Fig.1) is higher than that of the ball (low voltage part, which is shown in Fig.1); take the insulator samples with the running time above 8 a as an example, the defect possibility of the socket is 93.9%, and the defect possibility of the ball is 78.8%; high voltage part of end-fittings is vulnerable to electric field, which is the cause of the high defect probability.

Table 1 Sealing Performances of the End-fittings

Sample number Running time/a The number of non-defective insulators Ball Socket Rod The point and number of the sealing defect 33 >8 31 26 23 0 40 5~8 28 22 13 12 27 <5 6 3 2 21

Fig.1 The sketch of the composite insulator

2.2 Test analysis of the mechanical load tolerance

According to the experimental method from IEC 61109-2008 [15], the mechanical load tolerance is tested by automatic extension testing system, which is analyzed as follows:

(1) The mechanical load is applied on the insulators,which increases gradually to 70% (147 kN) of the rated mechanical load (210 kN);

(2) After keeping 147 kN for 1 min, the load increases to 210 kN during the time of 60 ~ 90 s;

(3) If a breakdown happens, the failure point and stress magnitude are recorded; if the breakdown is not happening,the tolerance test on the samples shows that there is no defect in the insulators.

The results of the mechanical load tolerance are shown in Table 2 and Fig.2. The results show that the possibility of breakdown in the insulators increases with the running time, and the breakdown possibility is about 100% with the running time above 8 a, and the breakdown load is lower than 180 kN; the breakdown possibility of the socket is also higher than that of the ball; taking the insulator samples with the running time above 8 as an example,the breakdown possibility of the socket is 51.9%, and the possibility of the ball is 29.6%.

With long-term chemical corrosions in the polluted industrial areas, the sealing of the end-fittings decreases with the running time. The acidic water permeates into the interface between the rod and end-fitting, and fiberglass of the rod has an ion exchange with the invaded acidic water,which results in the crackle of the rod, the sealing defect in the end-fittings, etc.

Table 2 Mechanical results of composite insulators

Sample number Running time/a The number of non-defective insulators Ball Socket Rod The point and number of the sealing defect 27 >8 14 8 5 0 32 5~8 12 8 4 10 25 <5 3 1 0 21

3 Monitoring theory

With the mechanical test data of the composite insulators, the typical defects include the sealing defects,and the fiber glass rod crackle. The high voltage part of end-fitting is with high risk according to the discussion of the second section.

The monitor points of the insulator with FBG are shown in Fig.3. Three typical points of the insulators include the normal (Point A), the edge (Point B) and the end-fitting(Point C).

The strain and thermal parameters are two basic data sets for the diagnosis of insulators, but the following analytical method explains how to distinguish the difference between the two parameters.

Fig.2 Mechanical scatter grams of composite insulators 3 Monitoring theory

Fig.3 Three monitoring point of the insulators

3.1 Strain stress

The stress distribution of the insulator rod is a symmetrical force from the analysis of Liang [1], so FBG embedded in the rod is under non-uniform stress, especially Point B in Fig.3. It should be clear about effects of the strain stress on the wavelength shift of FBG embedded.FBG divided evenly into M units with enough size,each unit is considered with the same stress. The center wavelength of each unit is changed by external force,which can be described by an equation from [16]:

neffi is the effective refractivity of each FBG unit, λi is center wavelength, Λi is effective refractive cycle, and Δλi is wavelength variation. The center wavelength changes with force, and Δneffi changes as:

Ef is elastic module, νf is Poisson ratio, p11 and p12 are the elasto-optical coefficients, σri and σθi are radial and hoop strain of the unit, respectively. The axial strain stress σzi is,

εzi is the axial strain of each unit. According to equation (1),if Λi changes, this behavior can be expressed as:

Substituting equations (2), (3) and (4) into (1), reflection spectrum of FBG can be obtained by getting the wavelength shift of each unit. This paper has P11 = 0.121,P12 = 0.270, Ef = 70 GPa, vf = 0.17, neff = 1.467, Λ = 528.289 nm,Δnmax = 2×10-4, and M=50.

When FBG embedded into rod axis of an insulator,and horizontal axis is 435 ~ 500 mm at about 5 mm near the edge of end-fitting, as shown in Fig.3. Calculated results of reflective spectrum are shown in Fig.4 and 5,respectively.

Fig.4 The normal spectrum of FBG

Fig.5 The chirped spectrum of FBG

Non-uniform stress is compared in Fig.4 and 5, as well as center wavelength shift, chirped phenomenon appears in reflection spectrum of FBG near the edge. The reflection energy center is changed by external strain, and then the range of reflecting wavelength becomes larger with decreasing intensity of reflections, which is shown in Fig.5.Quantitative analysis is based on this chirped shift caused by the strain stress from the comparison.

3.2 Thermal stress

When FBG is without external stress, wavelength shift of FBG Δλd is only changed with thermal, which can be described by an equation from [16],

∂neff /∂T is thermal refractive index coefficient with ξ represent, (Δneff)ep is elasto-optical change caused by external thermal expansion, ∂neff /∂α is waveguide change caused by FBG variation, ∂Λ/∂T is linear coefficient of thermal expansion with α represent, the linear coefficient of the FBG axis is the same as that of the vertical axis according to [16], and ξ = 0.68×10-5, as well as α = 5.5×10-7 in this work. The unit wavelength shift ΔλB is,

The unit elasto-optical change is described as:

According to the report by Wen [17], the change of daily temperature is no more than 30 ℃ in China, so ΔT = 30 ℃is applied as the thermal change.

Three typical points of an insulator include the normal,the edge, and the end-fitting shown in Fig.3, which are shown by Fig.6. Fig.6A, 6B and 6C are thermal changes of the normal, the edge and the end-fitting, respectively.Supposing that temperature rises with ΔT = 30 ℃, the change of the normal is the smallest, as central wavelength rises from 1.5500 μm to 1.5506 μm (Fig.6A). The change of the edge is larger, being close to that of the end-fitting,where the central wavelength changes from 1.5500 μm to 1.5507 μm (Fig.6B). And the change of the end-fitting is the biggest, where central wavelength increases from 1.5500 μm to 1.5508 μm (Fig.6C). Comparison of the three points shows that wavelength shift of the end-fitting is the most susceptible to thermal variation for end-fitting, and it is a sealing ring of the rod. The comparison can provide quantitative analysis by thermal effects on wavelength shift of FBG.

Fig.6 The spectrums of FBG caused by thermal changes

3.3 Feasibility of monitor analysis

From the relationship analysis above, two impact parameters of thermal and strain can be obtained by FBG embedded, respectively. Thus, FBG sensors can be perfected with feasibility of the insulator status in realtime.

Selection and design of handpicked FBG sensors are the first step of designing this monitor system to get the appropriate physical parameters and in addition with implantation and encapsulation, which define thermal and stress integrated factors of FBG sensors. In this paper, the wavelength shift of FBG is smaller than 2 nm, and the spectrum width of applied light source is 40 nm. Other parameters of applied system are in accordance with the application of insulators.

4 Designs and calibration tests

4.1 Design flow

The design flow for online monitor feasibility of this system is shown in Fig.7. The intelligent section is FBG sensor embedded in the rod parts of insulators, and so the basis of the system applications is how to select and implant FBG as sensors.

Implantation of FBG is the foundation of the system.Fiber grating is gotten by ultraviolet-writing technology using Single-mode optical fiber, which is implanted into the edge of the rod in the process of rod production.Wavelength range of FBG is 1529 ~ 1570 nm, wavelength precision is 10 pm, and scan frequency is 1 Hz. Optical mediator is used to get the wavelength changes of FBG caused by temperature and stress.

As an important unit, the back-sensing system consists of a power supply, a light source, a communication channel, an optical mediator, and a dispersion element. The back-sensing system is installed in the substation, and the selection of devices in the system is also on the basis of the applications.

Fig.7 Design flows of the monitor system

Optical fiber composite overhead ground wire (OPGW)is applied as the transport unit for input and output signal data. By OPGW, optical signal is generated from the light source and sent into FBG sensors, while monitored data of FBG is transmitted into data processing unit. This method can solve the online headachy problems such as power supply and data communication.

The data processing units are the most essential institution and expert diagnosis unit is included. Being an important unit of this system, the diagnosis can estimate the data for determination based on standards and experience from specialist team, according to result data from this unit.

With the effect of the implantation craft and the practical production of the insulator with FBG, the calibration test is carried out to get the strain and temperature characteristics of FBG in the insulators made from factory lines, and then get the coefficients of temperature and stress change. There are 12 units in the sample: FBG1, FBG2 in the ball (low voltage part of the fitting), FBG3 ~ FBG10 in the normal rod, and FBG11 ~ FBG12 in the socket (high voltage part of end-fittings).

4.2 Strain calibration test

The strain effects on the central wavelength changes of FBG have been discussed in the second part of the third chapter. A sample of the insulator with FBG is used to get the quantitative relations between the strain and the central wavelength change by the strain stress calibration test, and the relation coefficients can be used to analyze the strain of the insulators in running status. For the test time is so short that the cross effects between strain and temperature can be neglected, the test results only consider the effects of the strain stress on the chirped FBG (as shown in Fig.5).

Strain calibration is tested by robotized extension testing system, which is analyzed as follows:

(1) The mechanical load is applied to the insulators,which increases gradually from 10 kN to 50 kN;

(2) Keep loads for above 10 min, and when the central wavelength variation range of FBG units is 10 pm/min,the load, the central wavelength, and the surrounding temperature are recorded;

(3) Release the load and finish the test. The surrounding temperature is kept at 32.9°C in this test.

The strain calibration result data are in Fig.8, from which the conclusions are drawn:

(a) The central wavelength change has a positive linear relationship with the strain, and the linearity is above 0.99.Besides, the difference is quite modest among the slope factors of the positive line in the FBG units;

(b) The slope factors are 32 ~ 50 pm/kN; the factors are about 32 pm/kN in the end-fittings, which are smaller than the factors of the units in the normal rod (about 49 pm/kN);

(c) According to the results, the average of the factors is 49 pm/kN, which can be considered as the axial stress coefficient in a batch of insulator. The linear relationship in this work can be described by:

where εzi is above 5 kN, C is a constant, which is related to the insulator and FBG. It can be used as a distinction mark of the insulators.

Fig.8 The results of the load tests

4.3 Thermal calibration test

Thermal calibration test is done to get quantitative relations between temperature and the central wavelength changes of FBG, and relationships are used to analyze the thermal change of the running insulators. During the test time, the insulator sample is under a state of strain freedom,the test results only consider the effects of temperature changes on the wavelength shift of FBG (Fig.6).

The thermal and humidity chamber (STH-1000,Eastsun, China) is applied in the thermal calibration test,and the experiment steps are as follows:

(1) Put the insulators in the chamber, then increase the temperature from -20°C to 4°C, 22°C and 54°C;

(2) Keep each temperature for about 8 hours, so that temperature inside the insulators equals outside;

(3) When the central wavelength variation of FBG units is 10 pm/min, the central wavelength and the surrounding temperature are recorded.

The thermal calibration result data are in Fig.8, and the conclusions are drawn as follows:

(1) The central wavelength shifts have positive linear changes caused by temperature, and linearity is also above 0.99, and the difference is also quite modest among the slope factors of the positive line in the FBG units;

(2) The slope factor is the thermal coefficient in the units, and the factors are 13.5 ~ 14.1 pm/℃; the factors are about 13.5 pm/kN in the end-fittings, which are also smaller than that of the units in the normal rod (about 13.8 pm/℃);

(3) According to the results, the average of the factors is 13.8 pm/kN, which can be considered as the thermal coefficient in the batch of the insulators. The linear relationship in this work can be described as follows:

where D is a constant that depends on the insulator and FBG, and it can be used as the thermal characteristic value of the insulators.

Fig.9 The results of the temperature tests

5 Field application tests

The filed application test of monitor system has been carried out in transmission lines, and the obtained thermal and stress changes in an insulator is shown in Fig.10 and Fig.11, respectively.

From Fig.10, it is concluded that the daily thermal distribution with maximum value from the insulator changes closely related to its environment. The thermal value in Point C is the highest because the end-fitting with Point C is the high voltage part near the transmission line,and corona discharge provides heat from end-fitting and conductors nearby. The temperature values in Point A and Point B are generally the same, which are easily influenced by environment.

The result data from Fig.11 show that the residual strength value in Point A is the highest, while the value in Point B is the second, and the smallest value is in Point C. With the impact by the strong gale from August 7th to August 9th of 2012, monitor value of August 7th declined radically, and it dropped to the lowest point with increased gale in August 8th, but the value increased slowly with the decrease of typhoon in August 9th. However, in general,the residual value of strength is 80% ~ 100% after the gale of August 12th.

Fig.10 Thermal distributions of the insulators

Fig.11 Residual strength distributions of the insulators

6 Conclusions

This paper puts forward the method of using FBG as monitors for parameters correlation with thermal and stress of the composite insulators. The important conclusions are as follows:

(1) Based on the mechanical test results, the typical defect points have been found by the random composite insulator samples from the industrial pollution area, and these points are the focus of the monitoring.

(2) The parameters of FBG correlated with the strain and thermal are the fundamental detections. The calculations show linear effects of the thermal/strain stress changes on the center wavelength shift of FBG.

(3) Based on the calculations, the paper introduces the design of the composite insulator with FBG implanted, and the calibration tests are done to get quantitative relations for fault diagnosis of composite insulator defects.

(4) With field application tests, this monitor system has detected stress and thermal changes of the insulator caused by a strong gale during August 7th ~ August 9th of 2012.

Acknowledgements

This work is supported by National High-tech Research and Development Program of China (863 Program)(2013AA030701), and Science and Technology Project of the State Grid Xinjiang Electric Power Corporation(5230DK15009L).