Chapter 1
Principles of Transformer Differential Protection and Existing Problem Analysis

1.1 Introduction

With the development of the electric power industry, large capacity power transformers are more and more widely applied in power systems. As the heart of the whole power system, the performance of the transformer directly affects the continuous and stable operation of the whole power system. In particular, once a modern transformer of large capacity, high voltage, high cost and complicated structure is destroyed by a fault, a series of problems will emerge, such as wide-ranging impact, difficult and lengthy maintenance, and great economic loss. Statistics show that during the years 2001-2005, the average correct operating rate of transformers 220 kV and above is only up to 79.97%, far below the correct operating rate of line protection (more than 99%).

Differential protection is one of the foremost protection schemes used in the power transformer. The theoretical foundation of differential protection is Kirchhoff's current law (KCL), which is applied successfully in the protection of transmission lines and generators. However, there are many problems when it is necessary to identify transformer internal faults under various complicated operation conditions [1]. From the perspective of an electric circuit, the transformer's primary and secondary windings cannot be treated as the same node, with the voltage on each side being unequal. Besides, the two sides are not physically linked. In terms of basic principle, transformer differential protection is based on the balance of the steady magnetic circuit. However, this balance will be destroyed during the transient process and can only be rebuilt after the transient process is finished. Therefore, many unfavourable factors need to be taken into account in the implementation of transformer differential protection:

• Matching and error of the current transformer (CT) ratio.
• Transformer tap change.
• Transfer error of the CT increases during the transient process of the external fault current.
• Single-phase earth fault on the transformer's high voltage side via high resistance.
• Inter-turn short circuit with outgoing current.
• The magnetizing inrush.

With respect to the scenarios listed above, solutions to the first five mainly rely on the features of the differential protection. The tripping resulting from the inrush current needs to be blocked for the purpose of preventing mal-operation. In this section, various problems in current differential protection principles and inrush current blocking schemes are firstly studied and discussed. Then, some novel principles for transformer main protection are proposed and analyzed. Simulation and dynamic tests are carried out to verify the validity and feasibility of the novel principles. By comparative research, the development route of the transformer main protection technology is given.

Compared with EHV (Extra High Voltage) power systems, the electromagnetic environment of UHV (Ultra High Voltage) systems is more complex. Meanwhile, the configuration and parameters of an UHV transformer differ from an EHV transformer. In this case, the preconditions of applying transformer differential protection correctly rest with the modelling of the UHV power transformer reasonably and appropriate analysis of corresponding electromagnetic transients. The autotransformer is the main type of UHV transformer. However, the model of the autotransformer is not available in most simulation software. An ordinary countermeasure is to replace the autotransformer by the common transformer when executing electromagnetic transient simulations. In this case, the effect of magnetic coupling can be included but the electric relationship between the primary side and the secondary side cannot be involved. One of the existing models adopts the flux linkage as the state variable and includes the nonlinearity of the transformer core. It is clear in terms of concept but too complex to perform in many cases. In contrast, a new transient simulation model of the three-phase autotransformer is described, in which the controlled voltage and current sources are developed with the modified damping trapezoidal method, which is engaged to form the synthetic simulation model. In this case, both the efficiency and the precision of simulations are improved. However, this type of model will be more reasonable if it takes into account the nonlinearity of magnetizing impedance. Furthermore, the electromagnetic transient simulations in the UHV electromagnetic environment are new challenges, especially when including the UHV transmission line with distributed parameters.

Differential protection is usually the main protection of most power transformers. The key problem for the differential protection is how to distinguish between the inrush caused by unwanted tripping or clearing the external fault and fault currents rapidly [2-4]. The traditional methods of identifying the inrush are based on the theories of second harmonic restraint and dead angle. The flux saturation point becomes lower with the improvement of iron materials. The percentage of the second harmonic in the three-phase inrush current is probably lower than 15% in the case of higher residual magnetism and initial fault current satisfying certain constrains; the lowest might be under 7% with the relative dead angle smaller than 30°. The transformer differential protection cannot avoid the mal-operation regardless of whether second harmonic restraint and dead angle based blocking schemes are adopted. The theory of identifying the inrush using currents and voltages faces the problem of low sensitivity because of the difficulty of acquiring precisely the parameters of transformers. On the other hand, if the percentage of the second harmonic within the fault current is greater than 15%, this will cause a time delay in tripping of the protection based on the second harmonic criterion. This is due to the long-distance distributed capacitance and series compensation capacitance resonance in the high voltage power systems. The percentage of the harmonic will be larger if the characteristic of CT is not good (easy to saturate) and the differential protection cannot operate with the restraint ratio of 15%. Therefore, it is necessary to find a new criterion to identify the inrush for optimizing the characteristic of the differential protection of the power transformers.

1.2 Fundamentals of Transformer Differential Protection

1.2.1 Transformer Faults

Transformers are used widely in a variety of applications, from small-size distribution transformers serving one or more users to very large units that are the essential parts of the bulk power system. Moreover, a power transformer has a variety of features, including tap changers, phase shifters, and multiple windings, which requires special consideration in the protective system design.

Transformer faults are categorized into two classes: external faults and internal faults.

External faults are those that occur outside the transformer. These hazards present stresses on the transformer that may be of concern and may shorten the transformer life. These faults include: overloads; overvoltage; underfrequency; and external system short circuits. Most of the foregoing conditions are often ignored in specifying transformer relay protection, depending on how critical the transformer is and its importance in the system. The exception is protection against overfluxing, which may be provided by devices called 'volts per hertz' relays that detect either high voltage or underfrequency, or both, and will disconnect the transformer if these quantities exceed a given limit, which is usually 1.1 per unit.

Internal faults are those that occur within the transformer protection zone. This classification includes not only faults within the transformer enclosure but also external faults that occur inside the current transformer (CT) locations. Transformer internal faults are divided into two classifications for discussion; incipient faults and active faults.

Incipient faults are those that develop slowly but which may develop into major faults if the cause is not detected and corrected. They are of three kinds-transformer overheating, overfluxing, or overpressure-and usually develop slowly, often in the form of a gradual deterioration of insulation due to some causes. This deterioration may eventually become serious enough to cause a major arcing fault that will be detected by protective relays. If the condition can be detected before major damage occurs, the needed repairs can often be made more quickly and the unit placed back into service without a prolonged outage. Major damage may require shipping the unit to a manufacturing site for extensive repair, which results in an extended outage period.

Active faults are caused by the breakdown in insulation or other components that create a sudden stress situation that requires prompt action to limit the damage and prevent further destructive action. They occur suddenly and usually require fast action by protective relays to disconnect the transformer from the power system and limit the damage to the unit. For the most part, these faults are short circuits in the transformer, but other difficulties can also be cited that require prompt action of some kind. The following classifications of active faults are considered:

1. Short circuits in Y-connected windings
   1. Grounded through a resistance
   2. Solidly grounded
   3. Ungrounded.
2. Short circuits in...