Saturation detection and compensation of CT secondary current distortion taking into account the typical power system structure change

Master's Thesis in Electrical Engineering

Power Orientation

Abstract

In this thesis, while investigating the saturation phenomenon in protective CTs, the problems related to detecting this phenomenon and compensating the CT secondary current have been discussed, and in order to solve the problems, methods have been introduced and the results obtained in The software environment and modeling have been compared.

In order to reveal the saturation phenomenon, methods based on:

third order derivative,

discrete wavelet transform,

morphology leading,

and mathematical morphology

has been used.

To compensate for the secondary tortuous current, methods:

minimum error square,

flow estimation Magnetization,

and artificial neural network considering the structural changes of the sample network (and as a result the change of short circuit strength at the CT installation site) have been used to train this network.

After implementing and comparing the mentioned methods, mathematical morphology and least square error methodology as the most suitable method to reveal the phenomenon of saturation and compensation CT secondary currents have been proposed.

In addition to the above-mentioned cases, it has been tried with changes in the method of applying mathematical morphology (for detection) and the least square error method (for compensation of secondary currents), to provide the possibility of using the mentioned methods in online conditions.

The characteristics of the CT core examined in this thesis are also based on a practical test on the core of a CT real is extracted and finally, the resulting model has been applied and investigated in a part of Iran's simulated network (in the EMTP-RV software environment).

Key words: current transformer, detection of CT saturation phenomenon, third order derivative, discrete wavelet transform, forward morphology, mathematical morphology, compensation of secondary tortuous current, least squares of error, estimation of magnetization current, artificial neural network

1-1-Introduction

According to the development of the industry and the extent and complexity of the power systems, the level of short circuit in the power system is increased, which has increased the role of protective relays and interface equipment in preventing damage to pressure equipment in power systems. These relays need to receive the correct information to function properly, and therefore, in case of distortion in the received signals, expecting the desired performance from them is considered pointless. Current transformer (CT) [1] is one of the most important elements as a relay interface, which is used to obtain a current signal corresponding to the primary current and with smaller amplitudes. Although CTs use iron cores to maximize the bonding flux between the primary and secondary windings (and minimize the leakage flux), they are prone to saturation due to the non-linearity of the magnetic core. At points above the knee of the magnetization curve, the core magnetic current will increase significantly for the initial current changes. Since the secondary current of CTs is obtained from the difference of the primary transformer current and the magnetization current, under saturation conditions, the secondary current with a constant ratio does not follow the primary current and in addition to increasing the conversion ratio error, distortion will appear in the output signal. When a fault occurs, due to the DC component of the fault current (which is usually not included in CT design), the saturation phenomenon will occur, and one of the ways to limit this effect is to use a CT with a higher rated specification or use special algorithms to correct this phenomenon. Since the use of CT with higher rated specifications is not economically viable, software compensation of the phenomenon of CT saturation in power systems is a suitable solution to the problem, which leads to cost reduction and increased reliability.Since the use of CT with higher rated specifications is not economically viable, software compensation of the phenomenon of CT saturation in power systems is a suitable solution to the problem, which will lead to cost reduction and increase the reliability of the power system; Especially, such an algorithm can be easily applied in the structure of numerical relays (as an information preprocessor). Therefore, the purpose of this project is to detect the saturation phenomenon and compensate for CT secondary current distortion using signal processing methods.

1-2-Overview of the work done

As mentioned, due to current transformer saturation, in addition to increasing the conversion ratio error, the output signal will also be distorted. In [1-3], the problems caused by the occurrence of saturation in current transformers are investigated.

In [4], a method for detecting saturation in current transformers is presented based on the fact that the current changes rapidly when the saturation starts. This method detects CT saturation due to a sudden decrease in the amount of current, but it is not very successful if an anti-aliasing low-pass filter is used. In [5] and [6], a method for detecting current transformer saturation based on the third-order derivative of the secondary current is presented. In these articles, the effect of the anti-aliasing low-pass filter is considered.

In [7], an algorithm for calculating the core flux from the secondary current and then its compensation is proposed. This algorithm calculates the core flux well and detects CT saturation in different conditions. However, in this method, the assumption that the residual flux is equal to zero at the beginning of the calculations is used, which is not a suitable assumption in real conditions.

Another method to reveal saturation by calculating the average error and variance of the current range is proposed in [8]. The error value is determined with the assumption that if a sinusoidal current is complete, the sum of that current with a coefficient of its second derivative must be zero. In [9], an impedance method is proposed to detect saturation in a current transformer for busbar differential protection. This method is based on the first-order differential equation of the impedance of the power system source at the relay location, and it uses the busbar voltage and the secondary current of the current transformer to calculate the impedance. Changes in this impedance are used to determine the state of the current transformer. Also discussed are the effects of residual flux in the core, size of magnetization inductance and different fault modes. In [10], a detection method using symmetrical components for differential protection is proposed. In [11], another method for detection using artificial neural network and genetic algorithm is proposed. In this method, a neural network is used to detect saturation and a genetic algorithm is used to find the optimal structure of the neural network in terms of the number of layers and the number of neurons in each layer. In [12], a new hybrid method using the second derivative of current transformer output current and zero crossing rule is presented.

In [13], a compensation method is proposed in which, after estimating the CT core magnetization current, this current is added to the measured secondary current, so that the secondary current is obtained. This algorithm works well for different fault and system conditions, but (as in [7]) it is based on the assumption that the residual flux is zero before the fault occurs. The algorithm proposed in [14] compensates for the distorted secondary current and the residual flux level does not have an adverse effect on it. This algorithm uses a second-order differential equation to detect the moment of saturation.

An alternative method is to use an artificial neural network to estimate a function that corrects the current transformer secondary current that is distorted by saturation. This method has been used in many articles [15-19]. Dependence on the secondary capacity of the current transformer, failure to consider all the factors that can affect the saturation, and the non-optimal structure of the neural network are some of the defects that can be seen in these articles. In [20], the artificial neural network, the number of neurons and layers of this network is optimized by genetic algorithm, is used to detect and compensate for saturation.