Novel Intelligence ANFIS Technique for Two-Area Hybrid Power System’s Load Frequency Regulation

. The main objective of Load Frequency Control (LFC) is to effectively manage the power output of an electric generator at a designated site, in order to maintain system frequency and tie-line loading within desired limits, in reaction to fluctuations. The adaptive neuro-fuzzy inference system (ANFIS) is a controller that integrates the beneficial features of neural networks and fuzzy networks. The comparative analysis of Artificial Neural Network (ANN), Adaptive Neuro-Fuzzy Inference System (ANFIS), and Proportional-Integral-Derivative (PID)-based methodologies demonstrates that the suggested ANFIS controller outperforms both the PID controller and the ANN controller in mitigating power and frequency deviations across many regions of a hybrid power system. Two systems are analysed and represented using mathematical models. The initial system comprises a thermal plant alongside photovoltaic (PV) grid-connected installations equipped with maximum power point trackers (MPPT). The second system comprises hydroelectric systems. The MATLAB/Simulink software is employed to conduct a comparative analysis of the outcomes produced by the controllers.


Introduction
Now days, electricity is crucial because more and more individuals require it.Changes in the system's operating point and disturbances influence the system's dynamic behaviour.In power plants, the quality of the electricity generated is contingent on the machine's capacity [1].The rate and intensity of the electricity must remain constant, as intended.Therefore, load frequency control is a crucial component of the power system if it is to deliver reliable electricity [2].The most important goal of load frequency control (LFC) issues is to keep the frequency and two-area tieline power within a reasonable range in order to manage fluctuations in demand and disturbances [2].
Failure of grid was caused by this load frequency control issue, as is common knowledge.This is due to the excessive consumption of grid-sourced electricity in addition to its production.It resulted in a blackout throughout virtually the entire region, affecting all traffic.Due to the ineffective control of conventional controllers and despite repeated warnings, certain loads continued to utilize an excessive amount of energy..A system of robust control is required to detect load changes and stabilize frequency deviations.[3].
Traditional PID controllers can provide control actions for a single operating state, but in the real world, the parameters vary over time.Therefore, it is difficult to configure the appropriate gains so that there is no frequency shift.Since this is the case, an automatic solution is required.Load frequency controllers' dynamic performance has been enhanced by utilising various control strategies [3].
The most famous and most often used type of load frequency controller is the traditional PID controller.The standard controller is easy to set up, but it makes a big difference in frequency [4].Most state feedback controllers that are meant to improve performance are based on the idea of linear optimum control.Fixed-gain controllers are made for "nominal operating conditions," but they don't work well in a wide range of working settings [5].Keep track of the working conditions and use the most recent parameters to figure out how to run the system so that it works at its best.For LFC to do the same job as a PID Controller, adaptive controllers with gain settings that change on their own have been proposed [6].
Even when both renewable energy sources and the load cause disturbances, the proposed controller will make the frequency more stable.A neuro-fuzzy system can use both neural networks and fuzzy systems.So, an ANFIS-based controller will respond to changes in the environment by changing the membership functions of the fuzzy controller.This makes the fuzzy controller more flexible and reliable.The proposed technique (ANFIS) improves the performance of the system by training the settings of the fuzzy logic controller [7].

Proposed Scheme
Adaptive Neural Fuzzy Inference System (ANFIS) for building LFC in multiconnected areas with energy storage systems, which may reduce both control time and frequency variation during active power system operation [8].Artificial neural networks and fuzzy reasoning come together to make neurofuzzy systems [9].This technology uses neural networks to learn and fuzzy logic to figure out what to do.Neural networks can learn from data, but they can't tell you how they come to their conclusions.Fuzzy systems have linguistic rules that can be understood, can make decisions with imperfect data, and are good at explaining their choices, but they can't instantly learn new rules.Due to these limits, hybrid intelligent systems have been created [10].
A neuro-fuzzy hybridization creates a hybrid intelligent system by mixing the reasoning style of fuzzy systems, which is similar to how humans think, with the way neural networks learn [11].
Most of the time, neurofuzzy systems are shown as three-layer feedforward neural networks.The first layer is for the variables that come in, the second (secret) layer is for the fuzzy rules, and the third layer is for the variables that come out [12].
Each rule of the form if x = Ai, then y = Bi, where Ai and Bi are fuzzy sets, 1 ≤  ≤ , can be interpreted as a training pattern for a multi-layer neural network [13].
Figure 1 depicts the analytical structure of the proposed ANFIS-based load frequency management, which consists of the defuzzification, knowledge base, neural network, and fuzzy logic elements [14] [15].(1) Where   Are power angles of equivalent machines of the two areas

Modelling of PV System
The equation describes the transfer function of a solar photovoltaic system that includes a PV system, MPP tracker, converter, and filter [16] [17] .
Figure 3 block diagram of PV system Figure 3 shows the power electronic system collaborate with the corresponding second-order transfer function to describe the PV system, Where K -gain of the PV system, L -Zero (positive value) in the transfer function, M and N are poles (positive values) in the transfer function.[18].
A chopper circuit with MPPT and an inverter(grid-side ) with a filter comprise the PV system [19].
describing complex nonlinear interactions and are ideally suited for classifying phenomena into predetermined categories [20].However, the precision of outputs is frequently constrained and does not permit error-free results; it only permits the minimization of as few errors as possible [21].In addition, the training period for a NN may be quite extensive.In addition, the training data must be meticulously selected to represent the entire range over which the projected changes in the various variables will occur [22].Fuzzy logic systems effectively handle the imprecision of inputs and outputs by representing them using fuzzy sets, enabling the development of system descriptions with a desired amount of detail and flexibility.[23].
Neural networks and fuzzy logic have the capability to define mathematical relationships among multiple variables within intricate dynamic processes, enabling the representation of varying degrees of influence.Additionally, they offer the ability to control nonlinear systems to a level that surpasses the capabilities of conventional linear control systems.[24].
This is the last step in obtaining the precise output value O.

Results and discussion
Some standard benchmark measures are being used to test the suggested plan.Table 1 shows how well the proposed method works with different topologies.This test shows that the modified method is more likely to work.Table 1 shows how long it takes for different techniques from the suggested scheme to settle down at 1000 MW.It can be said that the suggested method gives better results than other methods.So, the proposed method is used to balance the load frequency and power for the said Two-area system.
Figures 6-23 show how the system responds to a 10% step increase in demand for a two-area system.As the system overshoots are less, it is clear that there is a big difference between the performance of the different schemes and the performance of the proposed ANFIS scheme.
To represent the results of the simulation, the following scenarios are considered: Case 1: Figures 6 and 7 show how the system responds without an arrangement in place.From what is seen, the frequency changes, power changes, and tieline power cannot be adjusted.Case 3: Figures 10 and 11 depict the system response utilising a discrete PID controller, in which observations are made, the frequencies (F1 and F2) are settled at 6.42 and 6.41 seconds, the power deviations (P1 and P2) are settled at 6.52 and 6.51 seconds, and the tie line power is settled at 6.49 seconds.
Figure 10.Frequency variations in two area system with discrete PID controller Figure 11.Power variations in two area system with discrete PID controller Case 4: Figures 12 and 13 depict the system response using the Fo-PID controller.In the observations made, the frequencies (F1 and F2) are stabilised at 6.13 and 6.15 seconds, power deviations are stabilised at 6.23 and 6.22 seconds, and connect line power is stabilised at 6.  the system response using the NN system, in which observations have been made, the frequencies have been settled at 5.8 and 5.9 seconds, the power deviations have been settled at 5.66 and 5.79 seconds, and the tie line power has been settled at 5.8 seconds.

Conclusion
In this paper, a novel method is presented.The problem of demand frequency balancing has been solved by developing a linearized model for renewable (PV)-based two-area power systems.Results from simulations and system indices indicate that the developed ANFIS outperforms the alternatives.The following conclusions can be drawn from the above observations:  Several operating conditions and performance indices reveal that ANFIS achieves satisfactory results. The developed ANFIS benefits from the NN and FLC exploration properties. When human export knowledge is unreliable, the developed ANFIS can be used to generate mature membership functions and fuzzy rules based on training data. Different indices and setting times guarantee the superiority of the proposed method.
The future scope of this work will consist of applying the proposed scheme to large-scale power systems.

Figure 1
Figure 1 ANFIS-based load frequency regulation analytical framework 2.1Mathematical model for Proposed Scheme Figure 2 represents the block diagram of hybrid power system consists of PV system, thermal and hydro system.

Figure 2
Figure 2 block diagram of hybrid power system .Power transmitted from area -1 is given by

Figure 4
Figure 4 block diagram of PV system (dc-dc converter with MPPT) 2.3 Modelling of Adaptive-Neuro Fuzzy System Fuzzy logic and neural networks are two distinct methods for addressing uncertainty.Each has both benefits and drawbacks.Neural networks are capable of

Figure 8 .Figure 9 . 7 E3S
Figure 8. Frequency variations in two area system with PID controller

2 seconds. 8 E3SFigure 12 .Figure 13 .
Figure 11.Power variations in two area system with discrete PID controller Case 4: Figures12 and 13depict the system response using the Fo-PID controller.In the observations made, the frequencies (F1 and F2) are stabilised at 6.13 and 6.15 seconds, power deviations are stabilised at 6.23 and 6.22 seconds, and connect line power is stabilised at 6.2 seconds. depict

Figure 14 .Figure 15 .
Figure 14.Frequency variations in two area system with NNsystem

Figure 16 .Case 6 :
Figure 16.Frequency variations in two area system with FLC

Table . 1
Representation of settling time of various technologies