Automatic Reactor Hunting Avoidance during Power System Restoration

Abstract

Modern society is a complex system that combines a lot of variables and one of the most important is electric power supply. Nowadays, our society extremely depends on the secure and consistent electricity supplies, being the main support of current industrial, economical, social and information activity. Furthermore, this dependence is becoming stricter as the society progresses.
However, power systems are very complex systems, which are vulnerable to faults of different nature due to high number of variables. The frequent danger that threats the complete availability of electric power are the large scale blackouts, occurred in case of lightning, natural disasters, operator’s errors, technical faults and so on. A large scale blackout has a critic impact on society, since electrical energy is inaccessible in the electric power grid and thus the economy is forced to stop.
As far as Swedish power system is concerned, it is characterized by long high voltage transmission lines connecting North of Sweden, which has the large percentage of the power generation of the country, to central and southern parts of Sweden where the power consumption is concentrated. In consequence, due to this system topology, the risk of voltage collapse at the central or southern parts of the country is an inherent feature of the Swedish system.
The ideal situation for a reliable power supply would be to prevent the system from any blackout. owever, blackouts may occur and it is impossible to predict when. Hence, power system estoration must be considered as a critic issue for Transmission System Operators (TSOs). The
main objective of TSOs is to perform a power system restoration as fast and safely as possible, n order to minimize the duration of the blackout.
During the restoration after the blackout in Sweden and Denmark on 23 September 2003, a articular problem appeared and increased the restoration time, known as reactor hunting. This roject is focused on studying this fault: its causes and its effect, and also ways of avoiding it in rder to reduce the restoration process.
In Sweden, the restoration strategy starts with long transmission lines from North to central part, eing energized. This leads to high voltage at the end of transmission lines due to the Ferranti ffect. In order to reduce the voltage, shunt reactors are connected to he system. Shunt
reactors are devices that are used to control the voltage in transmission lines and are controlled y what is called Extreme Voltage Automatics (EVAs).
The EVAs have a tolerance band which defines the behaviour of the shunt reactor. The olerance band has an upper and lower voltage limits and the objective is to maintain the voltage evel within these limits. If the voltage is above the upper limit of the tolerance band, the EVAs
connect the device; and on the other hand disconnect it if the voltage level is below the lower imit.
It is important to note that, during restoration, the power system is weak. Thus, connecting the eactors will probably produce critic low voltage (below the lower limit), then the EVAs will turn ff the reactor again, returning the voltage to the high level. The reactor will herefore cyclically onnect and again disconnect.
II
This phenomenon is called reactor hunting, which produces large voltage fluctuations between igh and low voltages outside the tolerance band. Thus it has a negative effect on the power system and needs to be handled as quickly as possible in order to prevent this fluctuations and possible damages.
The conventional procedure for TSOs to avoid reactor hunting is to deactivate the automatics during restoration time. This leaves the shunts in manual operation, which leads to longer restoration process.
It is almost straightforward to imagine that an automatic method for reactor hunting avoidance will be faster than manual operation and the restoration time will suffer an important decrease, which will be beneficial in blackout restoration.
This project presents a proposal for real time automatic reactor hunting avoidance founded on the “Adaptive tolerance band” concept, which adapts the EVAs behaviour based on the network strength. This new control scheme uses short circuit capacity to predict the voltage drop after shunt reactor connection and then adjusts the lower limit of the tolerance band in order to maintain the voltage within the limits and prevent reactor hunting from happening.
The important thing of this methodology is that it predicts the decrease in voltage after shunt connection. If the prediction is accurate, the lower limit will be adjusted precisely so any other external alteration of the voltage still can be detected by the automatics. If the adjustment is not accurate, an unexpected voltage drop may be ignored. It also needs to set a value for the lower limit that can not be reached to prevent extreme low voltages.
Consequently, the voltage drop prediction allows to detect when the extreme low voltage will be encountered after shunt reactor connection, so it can be decided if carry on with the restoration, take other path or take additional measures to control the voltage.
The core of the project is to implement this technique in a program which is able to perform a restoration process from an initial blackout scenario avoiding reactor hunting in real time with the automatic adaptive tolerance band method. This program allows the user to select any
restoration path possible.
Since the adaptive tolerance band is based on the network strength, the program is designed to calculate the Short circuit power from the Thevenin impedance of the specific point of the network for each restoration stage. This Thevenin impedance is calculated from the bus
impedance matrix. In consequence, computer methods for network matrices are implemented in this project in order to construct and modify the bus impedance matrix representing the situation of the power system at each moment.
The bus impedance matrix represents the actual strength of every point of the system for any particular time, so it is useful for the purposes of this project and also for other applications, such as short circuit faults, protection device design, etc.
In order to test the viability of the idea presented, this project is done in computer and simulation environments. The NORDIC32 system, which is a simplified model of the Nordic electric power system is used for this project as the test model.
The software used for power system simulation is ARISTO. The real-time program for Reactor Hunting avoidance is developed in MATLAB programming environment and the communication between both is done by AMCX communication tool.
III
Simulations with two different restoration paths, starting at the same initial southern blackout scenario of the NORDIC32 system, will show the effectiveness of the method proposed in the project for reactor hunting avoidance and will show the reactor hunting phenomenon before and
after avoidance.
To sum up, this project, by studying one particular problem during system restoration (reactor hunting), makes a great view of the actual understanding of what an electric power system is, the importance of blackouts and system restoration; and also, some of the computer methods
used in power systems such as power system simulators or algorithms for network matrices.
This topic is known as Energy Management Systems (EMS) which has an extreme importance for TSOs around the world.

• author: Amor Alonso, Guillermo
• supervisor: Mohammad Reza Safari Tirtashi; Olof Samuelsson; Jörgen Svensson
• organization: Industrial Electrical Engineering and Automation
• course: EIE920 20161
• year: 2016
• type: H2 - Master's Degree (Two Years)
• subject: Technology and Engineering
• keywords: Power system restoration. Reactor hunting. Adaptive tolerance band. Real time simulations. Voltage control. Power system modelling. Network matrices. Short circuit power. ARISTO.
• publication/series: CODEN:LUTEDX/TEIE
• report number: 5372
• language: English
• id: 8873950
• date added to LUP: 2016-06-30 14:18:18
• date last changed: 2016-06-30 14:18:18