Current regulations in the Brazilian electricity sector prevent transmission and distribution utilities from deenergizing equipment to perform long-term preventive electrical tests. Faced with the prospect of huge financial penalties, utilities in this environment seek predictive monitoring techniques that can be applied to energized equipment.
As a result, transmission systems and distribution networks must have a high availability and reliability. Dissolved gas analysis (DGA) is a monitoring technique that meets all of these requirements when applied to high-voltage electrical equipment filled with insulating fluid. It is used internationally for detecting faults in transformers and switchgear in service, providing excellent results.
There are recommendations for specific applications of DGA, particularly the condition assessment of on-load tap changers (OLTC). In this component, the arc breaking occurs within the insulating oil between the OLTC contacts and resistors. This operation could interfere with the identification of actual faults and defects.
CEMIG Geração E Transmissão S.A. now has extensive experience in monitoring OLTCs based on the application of chromatographic DGA in the mineral insulating oil within the OLTC. While this practice was originally used on power transformers and voltage regulators in substations at 138 kV or lower — where the arc extinction occurs in the oil — the technique is now being applied without restrictions in all equipment monitoring programs for voltages ranging from 13.8 kV to 525 kV.
Tap-Changer Maintenance
The systematic preventive maintenance plans for OLTCs are primarily based on the manufacturer’s recommendations, generally linked to the number of tap-change operations or on time-based maintenance. However, there is not a perfect correlation between these two variables when considering the limit for the OLTC’s actual condition and the need for preventive maintenance.
This was confirmed in a 2015 survey by the International Council on Large Electric Systems (CIGRE), which showed almost 27% of major failures on transformers with voltages above 100 kV were attributable to the OLTC. These results led to the need for developing and using predictive monitoring techniques to establish more effective maintenance strategies for determining the optimal time for maintenance of OLTCs. This is a key factor in ensuring the equipment has a high availability combined with high reliability. According to research, predictive techniques based on DGA of the OLTC’s insulating mineral oil samples have achieved solid results, which is why CEMIG has been using them for more than a decade.
For utilities that already have in place a DGA monitoring program applied to the main tank of the transformer, they can focus on optimizing their in-house expertise in terms of trained field and laboratory staff who take and test the oil samples. This is a way to extend a predictive monitoring program to include OLTCs.
DGA Predictive Program
The DGA technique applied to OLTCs is a predictive monitoring tool. It is a fully consolidated technique that works for power transformers and shunt reactors of all voltage classes. The main characteristics of DGA condition monitoring for a given asset are as follows:
• Conducts monitoring with equipment in service
• Reduces the need for outages, increasing availability and reducing costs
• Enables better asset maintenance planning and scheduling
• Reduces equipment repair costs
• Serves as a decision-making tool following unplanned events because the equipment can be energized securely
• Helps to avoid financial losses in a demanding regulatory environment, particularly in Brazil, where any supply interruption causes loss of revenue, fines and, in extreme cases, loss of concession rights.
CEMIG only conducts DGA monitoring on OLTCs that have no direct contact between the insulating oil in the OLTC and the insulating oil in the main transformer tank. It performs DGA monitoring on two types of OLTC designs.
The first design is the tank type or tap changer installed in the transformer’s main tank within a separate oil compartment. The diverter switch or a selector switch (a combined diverter and selector) operates within this separate oil compartment. This switching principle is normally referred to as a high-speed resistor-type OLTC.
The second design is the compartment type in which the tap changer is mounted in an external compartment that may be connected to the main tank by a barrier board. The external compartment usually is either one compartment (a combined selector and diverter in one compartment) or two compartments (a separate diverter and selector).
For reactor-type tap changers, the compartment type is usually used. Resistor-type tap changers are used only in special applications like arc furnace transformers, where the separate compartments provide easier access to the components for maintenance.
Despite the variability of gas accumulation in the insulating oil, even under normal conditions in non-faulty equipment, it is possible to distinguish patterns that can be used to differentiate between normal and faulty behavior of the equipment.
The analytical methodology applied follows ASTM D3612 Method C (headspace), and currently, 548 OLTCs are being monitored. This number represents CEMIG’s entire fleet of OLTCs that have an oil volume that allows samples to be taken without inserting any other maintenance hazards. Not all gases are meaningful for DGA diagnoses applied to OLTCs when compared to the traditional main tank DGA.
Monitoring Frequency
DGA monitoring frequency is either semiannual or annual, depending on the equipment voltage or criticality. Whenever a suspicious variation in a key gas profile is observed, the time interval between the sampling is reduced to improve the probability of early fault detection. It is important the OLTC being monitored has sufficient oil volume, so samples can be taken without causing a maintenance hazard. A minimum volume of 150 L (39 gal) in the OLTC is recommended.
When the DGA monitoring program first started at CEMIG, the utility soon realized the accuracy of the results improved when a sufficient volume of DGA data was procured for common types of OLTCs with operational and criticality similarities. After analyzing DGA information for each unit, the gas profile characteristics were evident for a given OLTC family.
Case Studies
In one case, it was possible to identify the presence of internal electrical discharges in the tap-changer of a 25-MVA, 138-kV step-down transformer manufactured in 1984. The diagnosis was confirmed, the transformer was shut down and the OLTC inspected. The defect was a transition resistor that had been damaged. Subsequently, it was replaced.
In another case, the OLTC on a 25-MVA, 138-kV step-down transformer manufactured in 1986 was monitored closely using DGA because of its high concentration of C2H2 compared with other OLTCs in its family group with similar operating conditions.
Another case included a typical C2H4 gas anomaly being observed in a sample, so the utility enhanced its time-based maintenance. In another instance, the diagnosis criteria indicated the presence of thermal degradation in oil based on an abnormal hot spot. The transformer was shut down and the OLTC was inspected and removed. It was transferred to CEMIG’s workshop, where action was taken to rectify loose contacts that could have caused severe damage to the OLTC and possible transformer failure.
The use of OLTC where the current interruption occurs in a vacuum chamber is relatively new, so the DGA monitoring database for this type of OLTC is still limited. Also, as expected, this type of OLTC is designed to operate longer than oil types without the need for maintenance. To date, no cases of incipient defects because of DGA predictive monitoring on this type of OLTC have been identified.
Benefits and Expectations
Over the last five years, CEMIG’s OLTC monitoring program has identified defects on 19 OLTCs, of which 14 were on 25-MVA, 138-kV transformers, representing a potential financial saving of approximately R$6 million. In 2016, the average capital cost of replacing a 25-MVA, 138-kV transformer was R$200,000, whereas the value to repair one of the damaged transformers in CEMIG’s workshop was R$15,000. The financial benefits of CEMIG’s monitoring program have more than covered the revenue expenditure of staff costs on the monitoring program.
The DGA technique can be applied with high accuracy to detect internal defects in transformer OLTCs and voltage regulators with operating voltages from 13.8 kV to 525 kV. For utilities that have not established a large DGA database to analyze the relevant variations in the concentration of gases, sufficiently reliable technical literature is available on which to base condition assessments. However, for future enhancements, a reliable database and the use of in-house experienced, trained personnel is recommended.
Acknowledgement
The authors wish to acknowledge the support and technical advice given by Dayve José Vassalo and Mauricio Joviano Proença, also with CEMIG, in the preparation of this article. ♦
Adriana de Castro Passos Martins has a master’s degree in metallurgical and materials engineering and is the monitoring center technical coordinator at CEMIG Geração E Transmissão S.A., where he has been responsible for the chemical and materials laboratory facility since 1997. Martins is a member of CIGRE and the convener for CIGRE Brazil Study Committee D1 Materials and Emerging Technologies.
Costabile Di Sessa holds a degree in chemical engineering from Escola Politécnica da Univerdade de São Paulo in Brazil. He joined CEMIG Geração E Transmissão S.A. in 2006, and since then, has worked as a maintenance planning engineer on materials, chemistry and plant monitoring. Sessa specializes in DGA monitoring diagnosis, polychlorinated biphenyls-related issues and SF6 technical support. He has been a member of CIGRE since 2012.
Laís Martins Marques Chaves holds a degree in chemical engineering from Universidade Federal de Minas Gerais in Brazil. She has been a planning and maintenance engineer for CEMIG Geração E Transmissão S.A. since 2014, acting in the electrical sector equipment predictive monitoring. In 2015, Chaves was the main author of the paper “CEMIG’s Experience – The Association of Scanning Electron Microscopy and Energy Dispersive X-ray Spectrometry Applied to Fault Analysis on Electric System Equipment,” presented at XXIII SNPTEE, a CIGRE Brazil technical event.
