Energy Management For Renewable Hybrid System Based On Artificial Neural Networks (ANN)

Abstract

It has become necessary in recent years to use renewable energy sources, such as wind and solar energy, to meet the load demand. Because of their endless existence and environmentally friendly nature, they are useful for power generation. The management of energy in hybrid wind-solar power sources is the topic of this article. Using a common current source interface multiple-input DC-DC converter, the photovoltaic (PV) array, wind turbine, and battery storage are all connected. The power control between intermittent renewable energy generation, energy storage, and the grid is maintained by the artificial neural network (ANN) controller. For wind turbines to produce the most electricity, variable speed control can be used. For solar systems, the Maximum Power Point Tracking (MPPT) algorithm is used. By keeping the common DC voltage constant, the grid interface inverter sends the energy drawn from the PV array and wind turbine into the grid. Using MATLAB Simulink, the complete control strategy is subjected to a simulation study. The results of the simulation show the photovoltaic/wind hybrid system\'s dynamic behavior and control effectiveness.

Introduction

I. INTRODUCTION

The energy that is obtained from resources that are found in nature, such as sunshine, wind, tides, waves, etc., is known as renewable energy. A hybrid generation system is used to provide enough electricity to fulfill the rising load demand. In a hybrid system, various renewable energy sources are integrated. The proposed system is a hybrid one that combines solar and wind power. Opportunities for using wind and solar resources for the creation of electric power have been made possible by advanced wind turbines and photovoltaic generating technology. They exhibit erratic, unpredictable conduct. Using both wind and solar energy effectively, the wind and solar integrated power supply system provides a fair source of energy.

Multiple control techniques are employed in a hybrid energy system for efficient and smooth power transfer. The type of converter used in the system, which is connected between the energy supply and the loads, determines the microgrid configuration. In order to regulate the energy among various resources, a control algorithm is needed [1]. It is applied in a setting with real-time control. The stand-alone hybrid system is employed to provide power to remote places connected to a flimsy grid. Depending on the weather, a hybrid energy-generating system is employed to satisfy the power requirement. The battery, which consists of lead-acid batteries linked in a series or parallel array, is used to store energy. The overuse of power compromises system dependability by causing an overvoltage on the DC bus. The adaptive control approach is utilized to mitigate reliability and over-voltage problems during excessive generating [2]. By keeping an eye on the battery's level of charge and the dc bus voltage, it is archived (SOC).

The processing and communication requirements are reduced by the energy management system-based multi-agent system (MAS). Middle-level and upper-level agents are its two agents [3]. The system is simple to maintain and regulate. The extraction of the load current and switching of the utility grid-side converter (UGC) is accomplished using an adaptive control technique based on least mean mixed norms (LMMN) [7]. It is used to lower the harmonic content of grid currents, enhancing the system's power quality.

In the hybrid system, various controller types are employed. PID controllers are the ones that are most frequently utilized. In fractional-order power generation and energy storage systems, frequency deviation is controlled using a fractional-order proportional-integral-derivative (FOPID) controller [4]. The output response oscillations are decreased, and the transient response is improved using a PID controller. The frequency deviation's settling time is shortened using a PID controller [11]. By using the wind turbine to drive the permanent magnet synchronous generator (PMSG), a gearbox setup that needs routine maintenance is avoided. A hysteresis controller is part of the hybrid wind-driven PMSG-PV system. It has excellent efficiency and a high-power factor. In order to get the most power out of both sources, PV and PSMG work together to create the current command [9].

Both the PV system and the wind energy conversion system employ artificial neural networks. It helps renewable energy sources perform as well as possible. The input layer, the hidden layer, and the output layer are all parts of an ANN's architecture [12]. The MPPT controller uses an RBFN-based Radial Basis Function Network (RBFN) to track maximum power from both solar and wind energy systems [5]. The two MPPT algorithms are combined to create a single RFBN-based MPPT controller that has been updated. As a result, the system became less complicated. Other neuro-fuzzy models can't match the speed of the adaptive neuro-fuzzy inference system (ANFIS). The three-phase inverter is subject to ANFIS-based regulation in order to manage the output power sent to the grid. It features two controllers: an active power controller and a reactive power controller [6].

Due to the resistance of time and place, it is exceedingly challenging to utilize solar and wind energy in all weather conditions alone through a solar system or a wind system. In order to address this need, a system based on renewable resources is also required, and wind/solar hybrid systems with battery storage can do so. This paper's main goal is to integrate hybrid solar and wind renewable sources to meet load power demand. The Luo Converter is used to increase efficiency and obtain significant voltage gains [8]. Low voltage ripple is provided by a Luo converter combined with great power density. The system's power is managed via an Artificial Neural Network. It is steady and dependable. By utilizing ANN, this system has low power and cost dissipation [10].

This paper is organized as follows. Section II introduces the proposed system Architecture and Modeling of Proposed architecture. In section III the artificial neural networks described. In section IV simulation and results Analysis of Proposed system. Finally, the summary of the proposed system is concluded in section V.

II. PROPOSED SYSTEM ARCHITECTURE

Photovoltaic (PV) arrays, wind turbines, multiple input DC/DC converters, and PWM inverters make up the proposed Artificial Neural Network-based energy management for the hybrid renewable system. Fig. 1 depicts the block diagram of the proposed ANN-controlled PV/wind hybrid generating system. Electricity is produced from solar energy by photovoltaic cells. Higher voltages, currents, and power levels are produced by connecting photovoltaic cells in series or parallel. Direct current is the current generated by the PV system (DC). When solar panels are utilized to generate power, there are no emissions that result in the greenhouse effect. Solar-generated electricity is crucial for the transition to the generation of clean energy. To track the maximum power generated by the PV system, MPPT technology is employed.

The wind is utilized to produce current through the wind turbine generator (WTG). The kinetic energy is transformed into mechanical energy using a wind turbine. Electrical energy is created by a generator using mechanical energy. Wind turbines use alternating currents as their source of current. Wind turbine output is sent to a rectifier, which converts AC to DC.

The multi-input DC-DC converter is supplied with energy from the wind and solar panels as input. As a DC-DC converter, the Luo converter is employed here. On the output side, it is used to step up the voltage. To increase the system's efficiency and achieve high voltage gain, a Luo converter is employed. A pulse-width modulation (PWM) inverter is used to convert the DC output of the Luo converter to AC. The inverter's PWM approach produces a constant output voltage regardless of the load. Without any dissipation, it regulates the amount of power given to the load.

A.  PV System Modelling

Fig. 2 depicts the PV cell's comparable circuit. A current source with a diode connected in parallel to it models a perfect solar cell. Nevertheless, no solar cell is perfect, so the model has series and shunt resistance incorporated, as seen in the picture. Rs is the very low value of the series resistance. The comparable shunt resistance, or Rp, has a very high value.

B. MPPT

The MPPT algorithm is based on the computation of the PV power and the power change by sampling both the PV current and voltage in the Perturb and Observe (P&O) approach. The tracker works by regularly increasing or decreasing the voltage of the solar array. While using instantaneous PV array voltage and current, the technique functions as long as sampling takes place just once during each switching cycle.

Up until the MPP is reached, the procedure is periodically repeated. After that, the system oscillates about the MPP. By decreasing the perturbation step size, the oscillation can be decreased. The MPPT, however, is slower with a smaller perturbation size. The flowchart for the traditional P&O technique is shown in Fig. 3.

The wind turbine, which gathers wind from the atmosphere and transforms it into mechanical energy, is the main component of the wind energy conversion system. This mechanical energy is used as input by the generator. This mechanical energy is transformed into electrical energy via the generator. The three-phase complete converter converts the three-phase AC produced by the generator into DC. After the DC bus, the inverter receives the DC.

D. LUO Converter

Fig.4 depicts the schematic circuit for the Luo converter. S is the power switch, and D is the freewheeling diode in the circuit. This LUO Converter is a multiple-input DC-DC step-up converter.

F. Three-Phase Spwm Inverter

Fig. 8 represents the three-phase voltage source inverter. High-power applications usually employ three-phase inverters. To form a three-phase voltage source inverter, the three single-phase half-bridge inverters must be linked in parallel [14]. The inverter comprises three phase legs, and each one consists of two switches and a body diode and is supplied by a constant DC voltage. In the SPWM control, the inverter switches are regulated by contrasting a triangular carrier signal with a sinusoidal modulating signal. The triangular carrier signal determines the inverter's switching frequency, while the sinusoidal modulating signal selects the required fundamental frequency of the inverter output. The modulation frequency ratio is the ratio of the triangular carrier signal to the sinusoidal modulating signal.

The intention of the three-phase inverters is to drive three-phase devices. Three sinusoidal signals 0 deg, 120 deg, and 240 deg are incorporated into the three-phase design at distinct phase angles. To produce positive and negative PWM pulses for the inverter, which should be fed to the switching devices, these three sinusoidal waveforms are compared with a positive and a negative carrier waveform [15]. The three legs of the switching devices should acquire the three distinct positive and negative pulses that make up the three phases. A scope is employed to monitor the switching pulses of all the switching device legs. Each switch is offered alternate switching of a positive and a negative signal at periodic intervals thanks to the organization of all the switching signals.

III. ARTIFICIAL NEURON NETWORK (ANN)

A. Artificial Neuron

ANNs are composed of artificial neurons which are conceptually derived from biological neurons. Each artificial neuron has inputs and produces a single output which can be sent to multiple other neurons. The inputs can be the feature values of a sample of external data, such as images or documents, or they can be the outputs of other neurons. The outputs of the final output neurons of the neural net accomplish the task, such as recognizing an object in an image. To find the output of the neuron we take the weighted sum of all the inputs, weighted by the weights of the connections from the inputs to the neuron. We add a bias term to this sum. This weighted sum is sometimes called activation. This weighted sum is then passed through a (usually nonlinear) activation function to produce the output. The initial inputs are external data, such as images and documents. The ultimate outputs accomplish the task, such as recognizing an object in an image.

B. Multilayer Neural Network

Artificial neural networks (ANNs), also known as neural networks (NNs) or neural nets, are computer architectures that draw inspiration from the organic neural networks found in animal minds. Artificial neurons, which are a set of interconnected units or nodes that vaguely resemble the neurons in a real brain, are the foundation of an ANN. Like the synapses in a biological brain, each link has the ability to send communication to neighboring neurons. An artificial neuron can signal neurons that are linked to it after processing impulses that are sent to it [16]. The output of each neuron is calculated by some non-linear function of the total of its inputs, and the "signal" at a connection is a real integer. Edges refer to the links.

The weight of neurons and edges usually changes as learning progresses. The weight alters a connection's transmission power by increasing or decreasing it. Neurons may have a cutoff and only send a signal if the combined signal passes it. Neurons may have a cutoff and only send a signal if the combined signal passes. Neurons frequently group together into layers. Different layers may alter their sources in distinct ways. Signals move through the levels, potentially more than once, from the first layer (the input layer) to the last layer (the output layer).

IV. SIMULATION

MATLAB/Simulink software is used to investigate the circuit behavior of the proposed Artificial Neural Network controlled solar-wind hybrid system. The simulation model of the hybrid system with Artificial Neural Networks is depicted in Figure 11.

The suggested system's simulation model is made up of a PV module, a wind module, a battery, a Luo converter, and an Artificial Neural Network. The photovoltaic system produces a DC current. PV may produce its most electricity when the MPPT technology is used. The wind turbine's output is in the form of AC. AC-to-DC conversion takes place in a rectifier. The DC-DC Luo converter receives energy from PV and wind as input. Currently, the Luo converter produces DC, which a PWM inverter converts to AC. The LC filter reduces the output signal's ripples and harmonics. The electric grid or AC loads are fed with the output.

The power management between renewable energy generation, energy storage, and grid is integrated by ANN Controller. In the ANN having inputs, hidden layers and outputs. The renewable energies and load are the input of ANN and the output of the ANN is Battery. The error (e) is used to adjust the neural network parameters. NN training involves supplying different inputs to the plant and teaching the network to map the corresponding outputs back to the plant inputs. The first stage of model predictive control is to train a NN to represent the forward dynamics of the plant. The prediction error between the plant output and the NN output is used as the NN training signal. The neural network plant model uses previous inputs and previous plant outputs to predict future values of the plant output.

A DC voltage of 75 volts is supplied by the solar module. Wind turbines provide an output voltage of 110 volts. When power is produced, the battery gets a 120-volt charge. The Luo converter combines the voltages from the sun and the wind. In order to convert DC to AC, the Luo converted the DC voltage and fed it to the inverter. The suggested ANN-controlled hybrid system produces 230 volts of final AC power. Figures 12, 13, 14, and 15 demonstrate the simulation results for the suggested ANN controlled solar/wind hybrid system.

Conclusion

In this Paper, an artificial neural network (ANN) controller-based hybrid solar/wind system is suggested for energy management. In the suggested hybrid system, the MPPT algorithm\'s efficiency is evaluated. The Luo converter is employed to maintain and adjust the DC link voltage. The high-frequency current harmonics in the wind generator could be reduced using the Luo converter. Power density and voltage gain are increased. The hybrid system\'s utilization of an artificial neural network (ANN) controller gives it reduces harmonics distraction and improved energy management. As a result, the overall performance of the hybrid system enhances operational effectiveness, quality, and system availability for power generation. According to simulation findings from MATLAB/Simulink, the suggested hybrid system is a realistic option for producing continuous electrical energy, particularly in rural locations.

References

[1] Wei, C., Zhang, Z., Qiao, W., & Qu, L., (2016). An Adaptive Network-Based Reinforcement Learning Method for MPPT Control of PMSG Wind Energy Conversion Systems. IEEE Transactions on Power Electronics, 31 (11), pp. 7837-7848. [2] Janvier Umuahia, Yuezhi Zhang, Shuang Zhao and H. Alan Mantooth, “An Adaptive Control Strategy for Power Balance and the Intermittency Mitigation in Battery-PV Energy System at Residential DC Microgrid Level”, IEEE Applied Power Electronics Conference and Exposition, pp. 1341-1345, Mar 2017. [3] B. Roobian,” Field-test upscaling of multi-agent coordination in the electricity grid,” in electricity distribution, 2009 20th International Conference and Exhibition on, June 2009. [4] Kumail Nusrat, Hamid Reza Mansouri, and Hossein Saboori, “Fractional-order PID controller design of frequency deviation in a hybrid renewable energy generation and storage system”, IET Journals, CIRED, vol. 17, no. 1, pp. 1148-1152, Oct 2017. [5] K. Kumar, N. Ramesh Babu, and K. R Prabhu, “Design and Analysis of RBFN-Based Single MPPT Controller for Hybrid Solar and Wind Energy System”, IEEE Access, vol. 5, pp. 15308-15317, Aug 2017. [6] Roseto, A., Panela, M., Aranea, R., & Andreotti, A., (2019). A Neural Network Based Prediction System of Distributed Generation for the Management of Microgrids. IEEE Trans. Ind. Appl., 55 (6), pp. 7092-7102. [7] Farhaan Chishti, Shadab Murshid and Bhim Singh, “LMMN Based Adaptive Control for Power Quality Improvement of Grid Intertie Wind-PV System”, IEEE Transactions on Industrial Informatics, vol.15, no.9, pp. 4900-4912, Sep 2019. [8] Jiao Y P, Luo F L, and Zhu M 2011 Voltage-lift-type switched-inductor cells for enhancing dc-dc boost ability: principles and integrations in Luo converter IET Power Electronics 4(1), pp. 131- 142 [9] L. Xu, X. Runa, C. Mao, B. Zhang, and Y. Luo, “An improved optimizing method for wind solar-battery hybrid power system,” IEEE Trans. Sustainable Energy., vol. 4, no. 3, pp. 774785, Jul. 2013. [10] Li, X., Zhang, X., Lin, F., & Blasberg, F., (2021). Artificial-intelligence-based design (AI-D) for circuit parameters of power converters. IEEE Trans. Ind. Electron., Early Access, Doi: 10.1109/TIE.2021.3088377. [11] Anu Rekha Nayak and Manoj Kumar Maharana, “Tuning of PID Controller to Maintain Load Frequency for Hybrid Power System”, International Conference on Innovative Mechanisms for Industry Applications, pp. 24-28, July 2017. [12] Balakrishnan, Hari Krishnan Nellippallil, Aditi Kathpalia, Sneha Shu Saha, and Nithin Nagaraj. 2019. \"Chaos Net: A chaos based artificial neural network architecture for classification.\" Review of. Chaos: An Interdisciplinary Journal of Nonlinear Science 29 (11):113125. [13] Jinn-Chang Wu, Chia-Wei Chou, “A Solar Power Generation System With a Seven-Level Inverter,” [14] IEEE Trans. Power Electron., vol. 29, no. 7, pp. 3454-3462, Jul. 2014. [15] Jinn-Chang Wu, Chia-Wei Chou, “A Solar Power Generation System With a Seven-Level Inverter,” [16] IEEE Trans. Power Electron., vol. 29, no. 7, pp. 3454-3462, Jul. 2014. [17] Jinn-Chang Wu, Chia-Wei Chou, “A Solar Power Generation System With a Seven-Level Inverter,” [18] IEEE Trans. Power Electron., vol. 29, no. 7, pp. 3454-3462, Jul. 2014. [19] Jinn-Chang Wu, Chia-Wei Chou, “A Solar Power Generation System With a Seven-Level Inverter,” [20] IEEE Trans. Power Electron., vol. 29, no. 7, pp. 3454-3462, Jul. 2014. [21] Jinn-Chang Wu, Chia-Wei Chou, “A Solar Power Generation System With a Seven-Level Inverter,” IEEE Trans. Power Electron., vol. 29, no. 7, pp. 3454-3462, Jul. 2014. [22] Krishna Kumar Gupta, Shailendra Jain, “A Novel Multilevel Inverter Based on Switched DC Sources,” IEEE Trans. Ind. Electron., vol. 61, no. 7, pp. 3269-3278, Jul. 2014. [23] Asuka Tsunoda, Youhei Hinago and Hirotaka Koizumi, “Level- and Phase-Shifted PWM for Seven-Level Switched-Capacitor Inverter Using Series/Parallel Conversion,” IEEE Trans. Ind. Electron., vol. 61, no. 8, pp. 4011-4021, Aug. 2014. [24] Badia, M.; McCarthy, R. Data estimation methods for predicting temperatures of fruit in refrigerated containers. Biosyst. Eng. 2016,151, 251–272. [25] Katic, K.; Li, R.L.; Verharrt, J.; Zeiler, W. Neural network based predictive control of personalized heatingsystems. Energy Build. 2018,174, 199–213. [26] Li, X.M.; Zhao, T.Y.; Zhang, J.L.; Chen, T.T. Predication control for indoor temperature time-delay usingElman neural network in variable air volume system.

Copyright

Copyright © 2023 M. Nagendra Babu, P. Pavan, K. Rama Mohan, G. Raja Rao, K. Srinivasa Rao, G. Tejaswi. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.