**4. Proposed protection system simulation and modelling**

In some cases, the generator starts to behave like a motor when the prime mover does not provide enough torque to keep the generator rotor rotating at the same frequency as the line of the parallel power source, and instead of giving power; it draws power from the parallel power source. Also, if the synchronisation ranges process rotates slowly, also both the loss of the alternator excitation. The governor is the fault of the original sender. Similarly, the generator will also extract the current from the source line [24]. When the rotating part of the generator fails, the generator stops generating electricity and starts drawing electricity from the parallel power source [25]. This situation may damage the drive

**207**

**Figure 2.**

*New High-Speed Directional Relay Based on Wireless Sensor Network for Smart Grid Protection*

unit and is not desirable. It should be detected as soon as a possible problem, and quickly disconnect the equipment from the parallel power supply, thereby protecting the generator from damage. In exceptional turbine cases, the power supply direction is changed from line to generator. It usually uses a directional protection relay to monitor the current flow and take appropriate action to prevent all outage case. The directional protection relay is working when the reverse power exceeds a certain percentage of the rated power output; it will trip the circuit breaker of the generator, disconnect the generator from the line under not normal circumstances, the relay setting is about 5% of the power generator [26]. The directional protection relay is located on the generator latch cabinet and is an integral part of the circuit breaker. The structure of the relay is designed to limit the reverse current flow depending on the amount of current and voltage between the two phase angles. If the line power is inverted, the current through the relay current coil will be inverted concerning the polarisation voltage and provide directional torque [27]. This technique compares the relative phase angle between the (current and

Typically, the phase angle is used to define the fault compares to the reference value. The voltage is usually applied as a reference amount. By comparing the operating voltage and current phase angle, can be inferred the fault occurs. Therefore, the fault current can be described with the phase relationship with the voltage line −90 for the forward fault, 90 for reverse fault. The relay wills response to the phase angle difference between the two quantities to come out trip signal [28]. In cases where optimal protection is required, Rogowski coil current sensors are used as CT and PT to avoid faults in conventional AC Transformers and must set a certain amount of delay during operation, to prevent power fluctuations, the transient effect during synchronisation. If the angle between the current vector and the voltage is Δ, the power flow is −900 < Δ < 900 [29]. Under normal conditions, the voltage overlaps with the current range is more significant than their non-overlapping interval. However, in the case of reversing energy flow, this overlap is reduced to a lower level. **Figure 3** shows this assembly and implementation. The low signal of the current and voltage of RC sensors changes to form a square wave having a value of "±1" and then multiplying these level signals to produce a positive number in the overlap interval, in the negative

The integration limit of the scheme is set to zero, therefore the integration of the load is perpetually <0 under normal conditions. However, in the opposite trend, the production condition system as a whole tends to decline until the threshold constant

numbers are generated in the non-overlap interval [30].

*According to the reverse power relay current with a voltage vector diagram.*

*DOI: http://dx.doi.org/10.5772/intechopen.85891*

voltage), as shown in **Figure 2**.

#### *New High-Speed Directional Relay Based on Wireless Sensor Network for Smart Grid Protection DOI: http://dx.doi.org/10.5772/intechopen.85891*

unit and is not desirable. It should be detected as soon as a possible problem, and quickly disconnect the equipment from the parallel power supply, thereby protecting the generator from damage. In exceptional turbine cases, the power supply direction is changed from line to generator. It usually uses a directional protection relay to monitor the current flow and take appropriate action to prevent all outage case. The directional protection relay is working when the reverse power exceeds a certain percentage of the rated power output; it will trip the circuit breaker of the generator, disconnect the generator from the line under not normal circumstances, the relay setting is about 5% of the power generator [26]. The directional protection relay is located on the generator latch cabinet and is an integral part of the circuit breaker. The structure of the relay is designed to limit the reverse current flow depending on the amount of current and voltage between the two phase angles. If the line power is inverted, the current through the relay current coil will be inverted concerning the polarisation voltage and provide directional torque [27]. This technique compares the relative phase angle between the (current and voltage), as shown in **Figure 2**.

Typically, the phase angle is used to define the fault compares to the reference value. The voltage is usually applied as a reference amount. By comparing the operating voltage and current phase angle, can be inferred the fault occurs. Therefore, the fault current can be described with the phase relationship with the voltage line −90 for the forward fault, 90 for reverse fault. The relay wills response to the phase angle difference between the two quantities to come out trip signal [28]. In cases where optimal protection is required, Rogowski coil current sensors are used as CT and PT to avoid faults in conventional AC Transformers and must set a certain amount of delay during operation, to prevent power fluctuations, the transient effect during synchronisation. If the angle between the current vector and the voltage is Δ, the power flow is −900 < Δ < 900 [29]. Under normal conditions, the voltage overlaps with the current range is more significant than their non-overlapping interval. However, in the case of reversing energy flow, this overlap is reduced to a lower level. **Figure 3** shows this assembly and implementation. The low signal of the current and voltage of RC sensors changes to form a square wave having a value of "±1" and then multiplying these level signals to produce a positive number in the overlap interval, in the negative numbers are generated in the non-overlap interval [30].

The integration limit of the scheme is set to zero, therefore the integration of the load is perpetually <0 under normal conditions. However, in the opposite trend, the production condition system as a whole tends to decline until the threshold constant

**Figure 2.** *According to the reverse power relay current with a voltage vector diagram.*

*Telecommunication Systems – Principles and Applications of Wireless-Optical Technologies*

**3.1 Wireless area network (WAN)**

has spread across a wide range of applications.

**3.2 Network home area (HAN)**

with secure communication [6, 10, 22].

**3.3 Neighbourhood area network (NAN)**

which the NAN network can be implemented [23].

**4. Proposed protection system simulation and modelling**

**Figure 1** shows the smart grid communication architecture, including the neighbourhood network (NANs), home LAN (HANs), wide area network (WANs), substation and data centre. These networks as follows will be displayed briefly [14–16].

It serves as backbones that help the power grid that provides communication between utility systems and substation systems. The can help prevent power outages by providing real-time information from the electricity grid. It supports real-time control and protection. This system is useful when dealing with unforeseen contingencies, and it is essential to avoid interruptions and failures [17–20]. This application helps in performing a generator process and provides support for large power systems. The main disadvantage of this kind of WLAN is the possibility of devices interfering at the same frequency is high. The network operates between 2.4 and 3.5 GHz. Among its advantages are low-cost equipment, the use of which

Some technologies introduced in the Home Network System are ZigBee, WLAN with PLC. The construction of a Building Area Network (BAN) is considered to be more complicated than the Home Area Networks (HAN). The HAN can be classified as a part of the customer network structure; HAN is often used by consumers in the housing and business sectors, using power tools to communicate [21]. It is a combination of connected devices, management software and dedicated LAN. HAN supports communication between smart meters and appliances used in homes, industries or buildings. It supports several other services, including Demand Response, pre-payment, real-time pricing and load control. The essential of the HAN communication system is to include low-cost, low power consumption

NAN is best described as a bridge between WAN and HAN and used in a NAN to collect data of points adjacent with the help of intelligent electronic devices (IEDs), which are widely deployed in the whole area. It is a two-way communication technology developed that give information about the control system for smart grids. Compared to WAN, the data rate is not high, and the transmission power is low for short-range transmission. WLAN, PLC and ZigBee are some of the techniques on

In some cases, the generator starts to behave like a motor when the prime mover does not provide enough torque to keep the generator rotor rotating at the same frequency as the line of the parallel power source, and instead of giving power; it draws power from the parallel power source. Also, if the synchronisation ranges process rotates slowly, also both the loss of the alternator excitation. The governor is the fault of the original sender. Similarly, the generator will also extract the current from the source line [24]. When the rotating part of the generator fails, the generator stops generating electricity and starts drawing electricity from the parallel power source [25]. This situation may damage the drive

**206**

**Figure 3.** *Diagram of implementing a directional component.*

**Figure 4.**

*Modelling of a directional relay component.*

is reached. In this situation, the constant is set to 0.01 and the select value based on the amount of reverse power [23]. The output of the reverse power relay (RPR) is transferred to a decision where the production is one for normal operation, zero for abnormal conditions, as displayed in **Figure 4**.

**Figure 5(a)** presents the 3ø current directions, cos ø, and power factor, and **Figure 5(b)** the same ideas of the P and Q expansion.

### **4.1 The adaptive neuro-fuzzy approach**

The selection of the membership function dramatically affects the quality of the fuzzy controller. Therefore, the method requires a more fuzzy logic controller. In this paper, a new method of neural networks is used to solve the adjustment problem of a fuzzy logic controller. We consider a dynamic system of multiple entrances, a single exit. The system is exported to the desired state of the control action can be described by the concept of the well-known "if-then" rule, where the input variables are first converted to their respective linguistic

**209**

follows:

for input and output.

**Figure 5.**

**Figure 6.**

incremental changes Δe(t) [31, 32]:

*Control method using adaptive neuro-fuzzy.*

*New High-Speed Directional Relay Based on Wireless Sensor Network for Smart Grid Protection*

variables, also known as fuzzification. The value of these rules certainly output. Then use defuzzification to convert the output to a precise value. For simplicity, we used a modified centre of area method, and the Triangle fuzzy set will be used

The linguistic form of the control rules is the basis of the designed fuzzy unit. It depends on the accuracy of the choice of parameters, which is the translation of the linguistic rules of the fuzzy set theory. The neural network (NN) is used to improve the selection of these parameters. In this scenario, the neural network is combined with the fuzzy logic unit. As shown in **Figure 6**, it uses the first fuzzy logic rule and then uses the neural network to generate the automatic adjustment output. References input [in (1)] related to the existing input [in (2)], product e(t), and

*Δ*e(t) = e(t) − e(t − 1) (1)

The proposed unit has two input factor gain measures of control, Ge and GΔe, and one scaling gain GΔu. The output-input scale factors are expressed as

*DOI: http://dx.doi.org/10.5772/intechopen.85891*

*(a) Quadrants of current/voltage. (b) Quadrants of a power.*

*New High-Speed Directional Relay Based on Wireless Sensor Network for Smart Grid Protection DOI: http://dx.doi.org/10.5772/intechopen.85891*

**Figure 5.**

*Telecommunication Systems – Principles and Applications of Wireless-Optical Technologies*

is reached. In this situation, the constant is set to 0.01 and the select value based on the amount of reverse power [23]. The output of the reverse power relay (RPR) is transferred to a decision where the production is one for normal operation, zero for

**Figure 5(a)** presents the 3ø current directions, cos ø, and power factor, and

The selection of the membership function dramatically affects the quality of the fuzzy controller. Therefore, the method requires a more fuzzy logic controller. In this paper, a new method of neural networks is used to solve the adjustment problem of a fuzzy logic controller. We consider a dynamic system of multiple entrances, a single exit. The system is exported to the desired state of the control action can be described by the concept of the well-known "if-then" rule, where the input variables are first converted to their respective linguistic

abnormal conditions, as displayed in **Figure 4**.

**4.1 The adaptive neuro-fuzzy approach**

*Modelling of a directional relay component.*

*Diagram of implementing a directional component.*

**Figure 5(b)** the same ideas of the P and Q expansion.

**208**

**Figure 3.**

**Figure 4.**

*(a) Quadrants of current/voltage. (b) Quadrants of a power.*

**Figure 6.** *Control method using adaptive neuro-fuzzy.*

variables, also known as fuzzification. The value of these rules certainly output. Then use defuzzification to convert the output to a precise value. For simplicity, we used a modified centre of area method, and the Triangle fuzzy set will be used for input and output.

The linguistic form of the control rules is the basis of the designed fuzzy unit. It depends on the accuracy of the choice of parameters, which is the translation of the linguistic rules of the fuzzy set theory. The neural network (NN) is used to improve the selection of these parameters. In this scenario, the neural network is combined with the fuzzy logic unit. As shown in **Figure 6**, it uses the first fuzzy logic rule and then uses the neural network to generate the automatic adjustment output. References input [in (1)] related to the existing input [in (2)], product e(t), and incremental changes Δe(t) [31, 32]:

$$
\Delta \mathbf{e}(\mathbf{t}) = \mathbf{e}(\mathbf{t}) - \mathbf{e}(\mathbf{t} - \mathbf{1}) \tag{1}
$$

The proposed unit has two input factor gain measures of control, Ge and GΔe, and one scaling gain GΔu. The output-input scale factors are expressed as follows:

$$
\Delta\_{cN}(t) = \Delta e(t).G\_{\Delta t},\tag{2}
$$

$$e\_N(\mathbf{t}) = e(\mathbf{t}).G\_\mathbf{t} \tag{3}$$

In the same eN and ΔeN scaling factor system to identify the fuzzy logic controller input signal product [33, 34]. Fuzzy logic controller output signal is ΔuN, it is the scale factor input. The neural network has two inputs, e(t) and Δe(t), and the neural network signal output α, which is used to fine-tune the product control of the operator. The output signal of the scale factor can be expressed by the formula:

$$
\Delta u(t) = \Delta u\_N(t) \text{tr}G\_{\Delta u} \tag{4}
$$

The output signal can be written as follows:

$$
\mu(t) = \Delta u(t) + u(t-1) \tag{5}
$$

**211**

**Table 1.** *Rules of FL.*

**Figure 8.**

to adjust the trip circuit:

*New High-Speed Directional Relay Based on Wireless Sensor Network for Smart Grid Protection*

*DOI: http://dx.doi.org/10.5772/intechopen.85891*

*f*(*x*) = <sup>1</sup> <sup>−</sup> *<sup>e</sup>*

*The description of the neural network compositions.*

*h*(*x*) = \_\_\_\_\_ <sup>1</sup>

The activation function neuron in the output layer is:

<sup>−</sup>*<sup>x</sup>* \_\_\_\_\_ 1 + *e*

1 + *e*

Control Unit based on the measured output signal *u*(t) from neuro-fuzzy circuit

θ*new* = θ*initial* − *k*.*u* (8)

where the θ initial is the initial switching output and k is a constant.

<sup>−</sup>*<sup>x</sup>* (6)

<sup>−</sup>*<sup>x</sup>* (7)

The results are displayed in **Figure 7**, which illustrates the specified fuzzy rules. We have selected fuzzy set and membership functions, **Table 1** summarizes the development of the rules used in this study [1].

Forming a neural network composed of three layers (two input layers, three hidden layers and one output layer). Neural network input (NN) is including the same number of output of fuzzy logic. The activated function has a value from −1 to +1 for the output signal, as shown in **Figure 8** [35–37].

**Figure 7.** *Fuzzy logic membership functions.*

*New High-Speed Directional Relay Based on Wireless Sensor Network for Smart Grid Protection DOI: http://dx.doi.org/10.5772/intechopen.85891*


**Table 1.** *Rules of FL.*

*Telecommunication Systems – Principles and Applications of Wireless-Optical Technologies*

*ΔeN*(*t*) = *e*(*t*).*G<sup>e</sup>*, (2)

*eN*(t) = *e*(*t*).*Ge* (3)

In the same eN and ΔeN scaling factor system to identify the fuzzy logic controller input signal product [33, 34]. Fuzzy logic controller output signal is ΔuN, it is the scale factor input. The neural network has two inputs, e(t) and Δe(t), and the neural network signal output α, which is used to fine-tune the product control of the operator. The output signal of the scale factor can be expressed by the

*u*(*t*) = *ΔuN*(*t*)α.*G<sup>u</sup>* (4)

*u*(*t*) = *u*(*t*) + *u*(*t* − 1) (5)

The results are displayed in **Figure 7**, which illustrates the specified fuzzy rules. We have selected fuzzy set and membership functions, **Table 1** summarizes the

Forming a neural network composed of three layers (two input layers, three hidden layers and one output layer). Neural network input (NN) is including the same number of output of fuzzy logic. The activated function has a value from −1 to +1

The output signal can be written as follows:

development of the rules used in this study [1].

for the output signal, as shown in **Figure 8** [35–37].

**210**

**Figure 7.**

*Fuzzy logic membership functions.*

formula:

#### **Figure 8.**

*The description of the neural network compositions.*

$$f(\infty) = \frac{1 - e^{-\varkappa}}{1 + e^{-\varkappa}} \tag{6}$$

The activation function neuron in the output layer is:

$$h(\infty) = \frac{1}{1 + e^{-\overline{\kappa}}} \tag{7}$$

Control Unit based on the measured output signal *u*(t) from neuro-fuzzy circuit to adjust the trip circuit:

$$
\Theta\_{new} = \Theta\_{initial} - k\omega t \tag{8}
$$

where the θ initial is the initial switching output and k is a constant.
