Generator Protection: Testing of protective functions ANSI 90, ANSI 46, and ANSI 59GN.

me . March 2013 Issue . Lessons Learned . Generator Protection

Generator Protection

Author: Torsten Schierz, OMICRON electronics Deutschland GmbH, Germany

Introduction

Generator protection systems are very complex systems with many different protective functions. The configuration of these systems depends on the rated power of the generator as well as on the power system structure, i.e. whether the generator is in busbar connection or connected to a unit transformer, as shown in Figure 4.  This article will describe solutions for testing selected protective functions effectively for the above connection configurations. As a starting point, experience recommends to structure the tests according to the following criteria:

Step 1:   Testing protective functions that only use voltage measurements (e.g. ANSI 59, ANSI 81, etc.)

Step 2:    Testing protective functions that only use current measurements (e.g. ANSI 50, ANSI 87, ANSI 46, etc.)

Step 3:   Testing protective functions that use both voltage and current measurements (e.g. ANSI 32, ANSI 90/40, etc.)

Test philosophy for overlapping protective functions

Generator protection systems can trip different breakers such as the network circuit breaker (CB), the generator CB, and the de-excitation CB. This so called trip command matrix depends on the activated protective functions, the technical philosophy as well as the power system structure.

Therefore it is recommended to test the correct behavior of each protection function. This is only possible without deactivating any protective functions during the tests.

Application examples

For the following protective functions, the relevant technical background will be explained and practical settings for testing will be derived:

• Testing the ANSI 90 protective function (underexcitation)
• Testing the ANSI 46 protective function (negative sequence / unbalanced load)
• Primary testing of the ANSI 59GN protective function (directional 90% stator ground fault)

Testing of ANSI 90

A synchronous generator always requires a sufficient DC voltage and thus a DC current flow through the excitation winding. This is necessary to maintain the synchronization to the connected power system. 

The generator capability diagram defines the limits of active and reactive power resulting from the physical parameters of the turbine and the generator, see Figure 2.

The underexcitation area is especially critical for the stability of the generator. In this area, the generator can lose its stability, e.g. as a result of a short-circuit in the connected power system or a malfunction of the automatic voltage regulator.

The protective function ANSI 90 protects the generator from asynchronous operation in case of these events.

Depending on the manufacturer of the protective devices, this function uses the impedance measurement or the admittance measurement.

Note: The impedance measurement method is not part of this article, because it is possible to transfer all necessary relay settings directly in the impedance plane of a distance relay (Figure 1).

On the one hand, for protection devices which use the admittance measurement method, the calculation of the relay settings is a lot easier than with the impedance measurement, because the admittance plane of the turbo generator can be used directly, see Figure 3.

On the other hand, testing the following characteristics in the admittance plane is rather complicated. Therefore, in the following, a more elegant method of testing admittance characteristics shall be introduced. The idea is to transform the admittance characteristics to the impedance plane to use automated distance test routines for testing the underexcitation protective function.

The admittance plane can be reproduced in the impedance plane using a mathematical transformation, as shown bellow.

The constant voltage  must be divided by the apparent power S. (equation 1)

                                                    (equation 1)

By using this formula any point in the admittance plane can be transformed to the impedance plane. However, for the automatic testing of the ANSI 90, it is desirable to transform the complete characteristic curve.

For this purpose the straight line equations must be inverted. According to Figure 5 and equation 2 the inversion of a straight line in the admittance plane results in a circle in the impedance plane.  

                                                                     (equation 2)

With   

Figure 6 shows the settings which are used in this example.

The following example shows the transformation of the trip time characteristic No. 1 to the impedance plane (equation 3) and the resulting impedance zone (Figure 7).

    

 (equation 3a)

                                                                 

 (equation 3b)

                                     

The same transformation applies to the trip time characteristic curves 2 and 3.  Figure 9 shows the results of this transformation.

Summary: With the transfer of the trip time characteristics from the admittance plane to the impedance plane, it is possible to use the same test philosophy as for a distance relay. In order to stabilize overlapping protective functions,such as the undervoltage protective function, it is required to use a constant test voltage, i.e. the rated voltage of the generator.

If the overcurrent protective function is active, its pick-up or trip may interfere for test points between 0 Ω and approximately -40 Ω in the impedance plane. For such cases the relay settings have to be checked.

Testing of ANSI 46

Unbalanced load conditions result in a positive and a negative sequence system, see Figure 8.

The negative sequence component rotates counter to the rotor movement and hence produces a flux which cuts the rotor at twice the rotational velocity. Thereby large currents with double frequency are inducted in the rotor causing severe heating.

According to the manufacturers it is possible to have different thermal trip time characteristics. This example is based on protective devices which use the ratio between the negative sequence current and the generator nominal current (I2/In). This function works with the current of only one side (side 1 or side 2) as well.

In multifunctional machine protection relays, all necessary protective functions, including differential protection are implemented in one device.

This can cause problems, because some protective functions may overlap during the test of the unbalanced load protection function.

One philosophy for testing this protective function is to deactivate the differential protective function before starting the test. The disadvantage of this method is that it is not possible to discover logic errors, e.g. in the trip command matrix or overlapping protective functions. Therefore it is recommended to test without deactivating any protective function.

Without a unit transformer in the protection zone, testing ANSI 46 with active differential protection is not too complicated. There is just a phase shift of 180° between the currents of side 1 and side 2.

Note:   The phase shift between side 1 and side 2 depends on the position of the CT starpoint grounding.

The situation is different, however, if there is a unit transformer in the same protection zone, as shown in Figure 11.

In this case, the vector group and the transformation ratio of the unit transformer must be considered. Also the different CT ratios will have an influence on the calculation of the test currents for side 1 and side 2.

The phasor diagrams in Figure 10 (a and b) display the phase shift of the test currents between side 1 and side 2 for a unit transformer with the vector group Ynd5.

The transformer ratio and the CT ratios are not considered in this diagram.

Note: For the positive sequence current the phase shift is 150° clockwise (transformer vector group 5 times 30°). The phase shift for the negative sequence current is 150° counter clockwise!

Summary:  As a conclusion, the physical behavior of the symmetrical components depends on the transformer vector group. It was shown that it is possible to test protective functions which only use one current measurement system (side 1 or side2) while the differential protective function is active.

Primary testing of ANSI 59GN

For generators with a maximum rated power lower than 50 MVA and busbar connection, the directional 90% stator ground fault protection (ANSI 59GN) is the standard protective function.

An alternative solution is the ground differential protective function with the displacement voltage as pick-up criterion and two cable-type transformers for ground current measurement, as shown in Figure 12.

Note: This solution is only possible, if the generator is connected via cable and the star point is accessible.

Secondary tests alone do not guarantee the correct function of the stator ground fault protection, because all settings were calculated based on theoretical grounding conditions.

It is therefore necessary to confirm the theoretical values with primary tests.

For this function the very first commissioning test is to check the secondary current transformer ground connection. Only one CT must be grounded.

The two primary tests (Figures 13 & 14) will check if the ground fault current is higher than the set pick-up value, and if the stability of the 90% stator ground fault protection function is ensured.

In this example, the settings of a 2.5 MVA generator (IEE > = 4 mA; Ven > = 4.9 V) were used.

Primary test with a ground fault inside the protection zone:  A ground electrode is connected to the generator terminal, the voltage regulator is deactivated and the trip command is blocked. When the generator runs at nominal speed the terminal voltage must be increased manually to the rated generator voltage. The next step is to measure the ground fault current (IEE) and the displacement voltage (Ven), as shown in Figure 13.

Primary test with a ground fault outside the protection zone (stability test): A ground electrode is connected to the feeder grounding point. The remaining test steps are the same as for the previous test. Figure 14 displays the test for this example.

Summary:  The comparison between the measured ground fault values for both primary tests confirms that the relay settings for the ground fault protection in this example are correct. With a ground fault which is located inside the protection zone, the flowing ground fault current (IEE) and the displacement voltage (Ven) are high enough to protect approx. 90% of the generator stator.

For the ground fault in the power system (outside the protection zone) this protective function is stable (no trip) and the difference between the parameterized pick-up value and the measured ground fault current (IEE) is high enough.  The theoretical value ∆I = 0A is not possible, because the cable-type transformers have different magnetizing characteristics.  If the comparison between the measured and the calculated values reveals potential malfunctions, the settings need to be adapted based on the measured values.

Conclusions

In this article, solutions for secondary and primary testing of the protective functions ANSI 90, ANSI 46 and ANSI 59GN were introduced.   These solutions have shown that secondary testing of complex multifunctional relays is possible even without deactivating overlapping protective functions. 

Furthermore, it was pointed out why the additional primary test of ANSI 59GN to verify the calculated settings is essential.

The presented approaches are illustrated in a way that allows for their application to similarly configured protective functions in relays of completely different manufacturers.  

Biographies

Biography:

Dr.-Ing. Torsten Schierz worked for 7 years at the University of Applied Sciences Zittau/Görlitz in the area of research and teaching. Since 1996 he has been an employee of OMICRON electronics Deutschland GmbH in the business fields Training, Commissioning and Technical Consulting. He has more than 16 years of experience in power system and rotating machine protection, especially in calculation and commissioning, as well as more than 24 years of experience in teaching electrical engineering. At present he is a Senior Consultant.

Torsten is member of the VDE (Federation of Electro Technology Electronics Information Technolog

The shifting asset management paradigm for Electrical Utilities

The shifting asset management paradigm

by Dr. Siri Varadan, UISOL

Ensuring that an asset performs to its full potential throughout its life is fundamental to effective asset management. Various factors make this objective difficult to accomplish, however. Tight budgets, vying priorities and a strict regulatory regime pose constraints that force utilities to do-more-with-less. Utilities are, as a result, shifting their thinking and moving to a paradigm where:

• Risk is no longer avoided, but managed;
• Costs are no longer minimized, but optimized; and
• Performance is no longer maximized, but adjusted to achieve thresholds.

In the context of electric utilities and ongoing smart grid efforts, this shifting paradigm means that asset management needs to be understood in terms of the following simple high-level questions: What work should be done? When and how do you do it correctly? While these questions are simple, they provoke thought on a variety of subjects throughout the asset management process shown in Figure 1. Going through the asset management process and focusing on the correct work answers the question: "Where should a utility invest its money to obtain the best return?"

Asset management, at a high level, addresses the following questions:

1. What assets does the utility own?
2. Where are these assets?
3. How important are these assets?
4. What is the condition of these assets?
5. What is the performance level of these assets?
6. Are these assets' conditions and performances satisfactory?
7. If not, should action be taken to restore the asset to its original performance or health?
8. If yes, what are the proper actions and how do you choose from a diverse set of actions so that corporate objectives, including customer satisfaction and regulatory approval, are satisfied?

While the first two questions almost sound trivial, they are fundamental to asset management and may be addressed by the implementation of a geographic information system (GIS) or an asset registry. A common thread in all of this is the availability and use of quality asset data.

Asset Criticality

Common sense dictates that the "squeaky wheel gets the grease." An asset that is of consequence should get more attention. In a recent asset management survey conducted by UISOL, released in May, utilities equated the word consequence to loss of revenue, system reliability and performance. Consequence may also be understood as the impact caused by the absence of an asset on the system, the customer, other assets and socio-economic factors. Risk is one of the better measures of asset criticality because it describes the impact of the failure of the asset by combining probability of asset failure and impact. Depending on the factors considered in its calculation, risk may take various forms—operational risk, environmental risk, public safety risk and so on.

Asset criticality in the electric utility industry is typically calculated per asset or by asset type and prioritized based on the asset's geographical and topological location. Figure 2 shows an example of asset prioritization for a utility with transmission assets. Value refers to the total sustainment expenditures and risk is a measure of the asset's loss consequence. To clarify, the loss of an asset in the category P1 has the greatest business impact.

Asset Health

Asset health is often considered subjective. All factors that determine asset health are not quantifiable and, hence, asset health is different from asset performance. Despite this, several efforts are used in the industry to quantify asset health. A score from 0 to 100 is sometimes used with the understanding that 0 means the asset is at end of life and requires immediate attention, repair or replacement. A score of 100 means the asset does not need attention for the next several years.

As a starting point, asset health can be conceived as a weighted average of several components, which is a measure of an attribute of the asset that could potentially lead to failure or result in a situation that could cause a failure condition. The asset health indicator should allow peer comparison, provide a sense of remaining life and indicate how soon intervention is required to avoid failure.

Identification of failure modes and the effects of these failure modes is important to health determination in reliability centered maintenance (RCM) analysis. Failure modes effects and criticality assessment (FMECA) focuses on evaluating a failure's impact. In doing so, a reliability engineer might focus on addressing failure modes that bear higher consequence. To eliminate human experts' subjective variations in asset health when selecting the weighting factors used to compute asset health indices, it is best to rely on statistical data and RCM studies that establish failure rates for each failure mode.

Asset Performance

Asset performance is a quantitative concept and correlated to asset health. The nature of the correlation, however, is a topic of further study. At a simple level, one might ask: How well is an asset performing with respect to its peers? The same question might be asked when comparing performance with other assets at other locations, perhaps owned and operated by neighboring utilities. As a result, it is important to understand benchmarking and utility best practices. It is also important to understand the role and nature of standards in evaluating asset performance.

Several measures for asset performance exist. These measures are mostly based on failure frequency and duration. Other metrics commonly used include restoration time, maintenance costs and time between failures. Financial metrics such as replacement costs, O&M costs and return on investment may also be included. Selecting and defining the metrics to use, and the logistics of data collection for calculating metrics are important when implementing an asset management project. It is difficult to calculate an individual asset's performance due to the lack of monitoring. It is possible, however, to make valid inferences about asset performance by considering data from a variety of sources. This is typically an area where data integration helps the most.

Data collected through online monitoring of electrical and non-electrical devices is common with smart grid. This new data can be used to assess asset health and performance when the systems are integrated effectively.

Asset Investment

Actions that restore problem assets to their original performance and health are necessary. These actions or projects could include asset maintenance, repairs, refurbishments or replacements. Each action has its pros and cons. Understanding the cost of these actions and their benefits over time is important when deciding which projects to implement. This science of decision making is at the core of asset investment planning (AIP).

Integrated AIP tools can assist in decisions making using a combination of objective functions, as well as constraints. AIP takes a list of projects and prioritizes them according to an established set of objectives. The rankings indicate projects' importance, their expected return and the time frame in which each project must be executed. AIP also provides information about risk associated with each project.

Asset management is a cradle-to-grave concept that requires careful asset planning, operations, maintenance, performance measurement and corrective actions to improve and maintain performance. Asking the right questions along each step of the asset management process is the best way to ensure goals are met. Internalizing the responses to each of these questions will enable a utility to transform to the new paradigm. A question can have more than one correct response. It is important to ensure that the answers work in concert to achieve asset management's overarching goal of identifying the correct work.

Present efforts at asset criticality, health and performance assessments combine data from various sources to provide quantifiable metrics that provide a sense of remaining life, when to take action and which action yields the most benefit. Correctly performing the work requires incorporation of best utility practices, tight integration of online monitoring, implementation of an asset management culture and personnel training. Leveraging smart grid efforts will be a key factor in the future of asset management.