24th International Conference & Exhibition on Electricity Distribution (CIRED) 12-15 June 2017 Session 1: Network components Managing on-load tap changer life cycle in tenaga nasional berhad (TNB) distribution power transformers Young Zaidey Yang Ghazali Asset Management Department, Tenaga Nasional Berhad, Petaling Jaya, Selangor, Malaysia E-mail: young@tnb.com.my ISSN 2515-0855 doi: 10.1049/oap-cired.2017.1308 www.ietdl.org Abstract: TNB has developed an asset management framework in accordance with ISO 55001:2014. One of the key elements of the framework is the asset life cycle management. This study presents TNB experience in managing the on-load tap changers (OLTC) over their complete lifecycle to improve performance and reliability. Using failure mode, effect and criticality analysis based on the previous failure data, lifecycle management strategies for OLTC have been identified. It covers the main activities over the lifecycle stages namely design, operation and. Adoption of new technology in the design, enhancement in condition assessment and effective implementation of conditionbased were amongst the strategies implemented. Finally, these strategies were validated through field evaluation and lifecycle cost analysis prior to the successful implementation throughout TNB. 1 Introduction Presently, there are 1260 power transformers rated up to 33 kv with on-load tap changer (OLTC), installed in TNB medium-voltage distribution network with capacity ranging from 7.5 to 30 MVA. About 70% of the power transformers and their respective OLTC are aged between 15 and 25 years. Out of the total population, 95% of the OLTC is of in-tank oil-immersed selector switch type with transition resistors or also known as oil-switch-type OLTC as shown in Fig. 1, which combines both functions of tap selector and diverter switch in one oil-filled compartment. The typical number of tap change operation of the OLTC in TNB distribution ranged between 2000 and 5000 tap change per year. Most power transformers studies indicated that the main cause of power transformers failures is the OLTC [1] since it has mechanical parts that are in constant movement. In TNB distribution network, there are seven cases involving permanent damage of OLTC since 2005. In most failure cases, it has caused prolonged power interruption, often due to severity of damage and timely replacement, as the result of unavailability of spares and incompatibility with the existing design. The root causes of failures are often attributed to degradation and carbonization of oil as well as contact problems as illustrated in Fig. 2. During normal tap change operation, switching arcs occur in oil due to the making and breaking of currents. These arcs cause degradation and carbonization that contaminate and reduce the dielectric strength of the oil [2]. The carbonization over time causes accumulation of carbon deposits on the surface of the fiberglass cylinder that resulted in the formation of electrical treeing (Fig. 3). Without proper, the electrical treeing develops into tracking and finally phase-to-phase arcing in OLTC. This is the common mode of failure found in most of the cases resulting in the permanent damage of the OLTC involved. On the other hand, coking or low conductivity film buildup was often observed on fixed and roller contacts during OLTC. This film consists of layer of pyrolytic carbon formation that bond to the oxide layer, which was formed as the result of surface oxidation of the contacts [2]. In some cases, coking has led to erosion or contact ware and pitted marks can be visibly observed. This occurs as the result of continuous overheating, as coking worsen over time due to increased contact resistance [3]. Prolonged condition of such phenomenon may result in excessive arcing that could trip the transformer protection (Fig. 4). Realising the needs to reduce the risk of OLTC failures, asset management strategies for OLTC have been proposed and implemented over its complete lifecycle. Thus, this paper describes TNB approach and experience in managing the lifecycle of OLTC in distribution power transformer to improve its performance and reliability. 2 Failure mode effect and criticality analysis (FMECA) FMECA as shown in Table 1 has been used to determine the most appropriate asset lifecycle management strategy for OLTC based on the actual findings on OLTC failures [4]. FMECA evaluation on other components of the OLTC such as motor drive and oil surge relay are not discussed in this paper since severity of failure of these components are still low. In addition, these components follow the standard design based on the manufacturer and type of OLTC. Furthermore, the operation and requirement of these components only involve visual inspection and functional check, which are very minimal. 3 OLTC lifecycle management strategies The complete lifecycle of an asset in accordance with asset management requirements of ISO 55001:2014 is illustrated in Fig. 5, where the activities at every stage of the lifecycle are also shown. Based on the FMECA in Table 1, TNB has identified the key activities at various lifecycle stages that can contribute to mitigate the causes of failure and improve the performance and reliability of power transformers as a whole. The lifecycle management strategies for OLTC based on these key activities are further discussed in the following subchapters. 3.1 Improving design and specification 3.1.1 Adoption of vacuum switch OLTC to mitigate the degradation and carbonization of oil in the OLTC 303
Fig. 1 In-tank oil-immersed selector switch type OLTC with transition resistors used in TNB distribution Fig. 2 Root cause analysis of typical OLTC failures in TNB distribution power transformers compartment as well as coking of arcing contacts: Degradation and carbonization of oil as described above are the most common causes of failure that has resulted in failure and permanent damage of the OLTC. In order to mitigate this problem, TNB has adopted the use of vacuum switch OLTC, shown in Fig. 6, that confines switching in interrupted vacuum bottles. As the results, this helps to prevent contamination of oil due to carbonization and hence lower the rate of oil degradation due to switching arcs. Furthermore, with the absence of oil inside the vacuum switch, formation of low conductivity film and deposition of carbon that lead to coking on the contacts surface will no longer occur. With the arc quenching property of the vacuum switch, contact erosion and thus contact wear is minimised and thus reduces costs. Prior to the adoption of the use of vacuum switch type OLTC in 2011, a field trial was conducted in 2007 to evaluate its performance. 3.1.2 Adoption of free self-dehydrating breather to mitigate the degradation of oil in the OLTC compartment: Degradation of oil in the OLTC is not only influenced by the switching arcs that occur due to making and braking of currents during on-load tap change operation, but also affected by the presence of moisture in the oil. Even with the use of vacuum switch type OLTC, moisture can still be presence due to leakage or in most cases due to lack of of the OLTC dehydrating breather (Fig. 7). In order to mitigate this problem, TNB has adopted the use of free self-dehydrating breather as depicted in Fig. 8. The breather has a heating element mounted within the container to heat the desiccant at selected intervals with temperature sensor to monitor the correct operation of the heater. A moisture sensor measure the humidity of the air to ensure only dehydrated air goes through the piping into the OLTC conservator. Both sensors are controlled by an electronic controller. The condensed moisture formed on the surface of the container will be expelled outwards by gravity. A field trial on the use of the self-dehydrating breather was conducted in 2007 prior to its adoption in 2011. 3.1.3 Lifecycle cost analysis: Prior to the adoption on the use of vacuum switch type OLTC together with free self-dehydrating breather, a lifecycle cost analysis (LCCA) over the expected transformer life span of 40 years is performed to validate and compare the net present value of the total lifecycle cost of using vacuum switch type OLTC together with self-dehydrating breather, and the oil switch type OLTC. The LCCA of the two systems shown in Table 2 reveals that the use of vacuum switch type OLTC together with self-dehydrating breather could generate saving of more than RM 600,000 a year for the entire transformer population, not including saving on the avoidance of loss revenue due to OLTC failure. 3.2 Enhancing methods for condition assessment Fig. 3 Formation of electrical treeing on the cylinder s surface transition resistors used in TNB distribution resulted in tracking and arcing between phases The adoption of vacuum switch OLTC is only applied for the new power transformers. Therefore, for the existing in-service power transformers, emphasis is given on the utilisation (operation) and lifecycle stages which are discussed below. 3.2.1 assessment using oil quality analysis to detect degradation of oil in OLTC: The implementation of the condition assessment for OLTC by means of oil quality analysis involved breakdown voltage and moisture content has started since 2007. Table 3 summarises the oil quality indicator limits used for OLTC. Fig. 4 Erosion and pitted marks on the roller contacts (left) as well as carbon buildup on the fixed contacts (right) were observed during 3.2.2 assessment using dissolved gas analysis (DGA) to detect coking of contacts: DGA has been applied to assess the condition of the OLTC since 2007. owever, the health condition of the OLTC then was not truly understood due to unavailable guide for interpretation of DGA results. In 2011, with the issue of IEEE Std. C57.139-2010 [5], a more enhanced condition assessment method using DGA data was formulated to 304
Table 1 FMECA based on common findings of OLTC failures Failure mode Failure effect S Failure cause Failure mechanism Failure consequence P C arcing between phases excessive arcing between contacts permanent tracking on fibreglass cylinder damage of metal parts damage of contact contact wear off degradation and carbonisation of oil tracking as the result of accumulation of carbon deposits on the surface of cylinder M coking of contacts overheating due to increased contact resistance transformer trip by protection relays transformer trip by protection relays M M Note: S = Severity of failure, P = Probability of failure, C = Criticality, = igh, M = Medium, L = Low. Fig. 5 Lifecycle activities of an asset Fig. 8 Use of free self-dehydrating breather Table 2 Comparison of the net present value of the total lifecycle cost over the life span of 40 years per OLTC system OLTC system Initial cost, RM Operation and cost, RM Total lifecycle cost, RM oil switch type OLTC and conventional breather vacuum switch type OLTC and self-dehydrating breather 160,814.00 69,652.23 230,466.23 196,145.00 15,198.46 211,343.46 Fig. 6 Vacuum switch type OLTC used in TNB distribution power transformers Table 3 Property indicators for OLTC oil quality analysis Limits Good Fair Poor Bad breakdown voltage, kv >50 40 50 30 39 <30 water content, ppm <15 15 30 31 45 >45 Fig. 7 OLTC Lack of of silica gel in the dehydrating breather of the interpret the health condition of the OLTC. The statistical model used follows the method described in Annex B of IEEE Std. C57.139-2010 to determine two types of condition indicator limits i.e. gas concentration and gas ratio limits. The model was initially based on DGA data up to the year 2011 taken from 503 units of oil-immersed selector switch type OLTC classified as ARAB type OLTC in accordance with IEEE Std. C57.139-2010 classification scheme. The gas concentration limits is defined in terms of a statistical outlier limit identifying extreme values of gas concentration suspected to be the results of faults or unusual stresses. The gases that are used as indicators to discriminate between normal and faulty conditions are C 2 2,C 2 4 and C 4 together with the sum of C 4,C 2 6 and C 2 4 called the total dissolved heating gases. The upper outlier limits U1, U2 and U3 were calculated for each set of gas concentration data. U3 is introduced to represent the extreme values requiring the highest attention (U3 = Q3 + 4.5IQR where Q3: third quartile, IQR: interquartile range). Table 4 summarises the gas concentration limits for OLTC. The gas ratios that are used as indicators to discriminate between normal and faulty conditions are C 2 4 /C 2 2 and TDG/C 2 2. The gas ratio values of non-faulty OLTC operating under normal 305
Table 4 OLTC gas concentration limits Concentration, ppm codes Generic C 4 C 2 4 C 2 2 TDG C U1 C 2083 C 3522 C 14598 C 7000 normal (1) U1 > C U2 2083 > C 3254 3522 > C 5569 14,598 > C 23004 7000 > C 11029 caution (2) U2 > C U3 3254 > C 4423 5569 > C 7617 23,004 > C 31410 11,029 > C 15058 warning (3) C >U3 C > 4423 C > 7617 > 31410 > 15058 danger (4) Note: C = gas concentration. Table 5 Ratio OLTC gas ratio limits Generic C 2 4 /C 2 2 TDG/C 2 2 codes R C 090 R 0.378 R 0.684 normal (1) C 090 > R C 095 0.378 > R 0.480 0.684 > R 0.843 caution (2) C 095 > R C 099 0.480 > R 0.743 0.843 > R 1.234 warning (3) R >C 099 R > 0.743 R > 1.234 danger (4) Note: R = gas ratio. Table 6 Results codes Interpretation of DGA results for OLTC indicator Interpretation R = 1 and C 2 good normal OLTC operation R = 1 and fair light coking or deterioration of arcing contacts, or unusually high frequency of tap change operation or high load current causing heating of transition resistors 1<R 3 and poor coking or increased deterioration of R = 4 and bad arcing contacts or severe deterioration Table 6 gives the overall interpretation to facilitate the understanding on the condition of the arcing contacts based on the results of the gas concentrations and ratios. Table 7 compares the outcome of the DGA results on 745 units of OLTC interpreted using the above method with the same DGA results using the Duval s triangle method for load tap changers [6]. 3.3 Effective implementation of condition-based Based on the enhanced condition assessment method above, an overall health condition of the OLTC is derived as shown in Table 8 that provides recommended actions for mitigations as tabulated in Table 9. Based on the health condition and recommended mitigating actions, more effective planning can be executed where work can now be prioritised according to the criticality of the OLTC condition as shown in the example given in Table 10. The assessment was validated through visual inspection of the OLTC condition. Fig. 9 shows typical observations found during. Validation of the assessment was also carried out by means of dynamic current measurement using dynamic winding resistance technique and it is discussed elsewhere [7]. Table 8 Overall health condition of the OLTC Table 7 Comparison on the number of problematic OLTCs detected using TNB condition assessment method derived from IEEE Std. C57.139-2010 and Duval s triangle for OLTC Fault identification TNB OLTC assessment method OLTC Duval s triangle method Overall health condition Oil quality analysis results Good Fair Poor Bad DGA results good 1 2 3 3 fair 2 2 3 3 poor 3 3 3 4 bad 3 3 4 4 total no. of possible faults detected no. of severe thermal faults with moderate or no. of thermal faults in progress with light coking or heating of transition resistors 59 55 41 22 18 33 Table 9 Recommended mitigating actions based on the overall health condition of the OLTC codes indicator Recommended actions conditions are described by the percentiles of the lognormal distribution representing the non-outlier gas ratios. The gas ratio limits calculated for the study are based on 90th (C 090 ), 95th (C 095 ), and 99th (C 099 ) percentiles of the lognormal distribution with significance level of 0.1, 0.05 and 0.01, respectively. Table 5 summarises the gas ratio limits for OLTC. It should be noted that the ratio limits in Table 5 are applied only when any of the gas concentration has reached warning limit or condition 3 in Table 4. 1 normal continue oil sampling at 12 month interval 2 caution continue oil sampling at 6 month interval 3 warning inspect for leaks and condition of silica gels. Conduct internal inspection on arcing contacts. Perform overhaul and replace oil 4 danger inspect for leaks and silica gels condition. Possible replacement of arcing contacts. Perform overhaul and replace oil 306
Table 10 Example of planning sheet based on the criticality of the OLTC health condition State Site TX no. TNB OLTC assessment Duval s triangle method DGA Oil quality ealth condition selangor commerce square T1 or severe selangor commerce T2 or severe square selangor proton T2 or severe selangor taman T2 coking or increased deterioration berjaya klang selangor labur bina T1 or severe selangor subang T2 or severe selangor jaya TC taman berjaya klang T1 coking or increased deterioration light contact or increased contact resistance light contact or increased contact resistance fault T3 ort2 in progress or severe arcing D2 (X3) light coking or increased contact resistance light coking or increased contact resistance bad bad warning have been identified for the implementation of the lifecycle strategy, which includes design, operation and of OLTC. Finally, validation through field evaluation and LCCA was performed to evaluate for technical and economic feasibility of the proposed technologies and methodology, prior to the successful adoption of all the strategies throughout TNB since 2011. Fig. 9 Surface erosion with pitted marks and coking of the roller contacts (left) and carbon deposits on the cylinder (right) were observed during OLTC at several sites 4 Conclusions Managing asset lifecycle is the main focus of asset management system starting from the creation of the asset, utilisation, up to its retirement. Thus, each aspect of the lifecycle activities must be strategised to optimise the usage of the asset and to strike a balance between cost, risk and performance of the asset. This paper has presented all possible mitigating actions that have become the OLTC asset management strategy at various stages of its lifecycle. Based on the FMECA, three main lifecycle activities 5 References 1 Jongen, R., Morshuis, P., Smit, J., et al.: A statistical approach to processing power transformer failure data. CIRED 19th Int. Conf., Paper 546, 2007 2 Erbrink, J.J., Gulski, E., Seitz, P.P.: Advanced on-site diagnosis of transformer OLTC. IEEE Int. Symp. on Electrical Insulation, 2008, pp. 252 256 3 Erbrink, J.J., Gulski, E., Smit, J.J.: assessment of OLTC using dynamic resistance measurements. Int. Conf. on V Engineering, 2010, pp. 433 436 4 BS 5760-5: Guide to failure modes, effects and criticality analysis (FMEA and FMECA) 5 IEEE Std C57.139: IEEE guide for dissolved gas analysis in transformer load tap changers, 2010 6 Duval, M.: The Duval triangle for load tap changers, non-mineral oils and low temperature faults in transformers, IEEE Electr. Insul. Mag., 2008, 24, (6), pp. 22 29 7 Shahril, M., anum, Y., Zaidey, Y., et al.: Diagnosis of OLTC via Duval triangle method and dynamic current measurement, Proc. Eng., 2013, 68, pp. 477 483 307