A Smart Microgrid System with Artificial Intelligence for Power-Sharing and Power Quality Improvement

Abstract

The widespread popularity of renewable and sustainable sources of energy such as solar and wind calls for the integration of renewable energy sources into electrical power grids for sustainable development. Microgrids minimize power quality issues in the main grid by linking with an active filter and furnishing reactive power compensation, harmonic mitigation, and load balancing at the point of common coupling. The reliability issues faced by standalone DC microgrids can be managed by interlinking microgrids with a power grid. An artificial intelligence-based Icos$\varphi$ control algorithm for power sharing and power quality improvement in smart microgrid systems is proposed here to render grid-integrated power systems more intelligent. The proposed controller considers various uncertainties caused by load variations, state of charge of the battery of microgrids, and power tariff based on the availability of power in microgrids. This paper presents a detailed analysis of the integration of wind and solar microgrids with the grid for dynamic power flow management in order to improve the power quality and to reduce the burden, thereby strengthening the central grid. A smart grid system with multiple smart microgrids coupled with a renewable energy source with tariff control and judicious power flow management was simulated for power-sharing and power quality improvement. A hardware prototype of the artificial intelligence-based Icos$\varphi$ control algorithm with nonlinear load was also implemented successfully. Furthermore, the economic viability was investigated to ensure the feasibility of the smart microgrid system with the proposed controller design for power flow management and power quality improvement.

1. Introduction

Conventional sources of energy are hampered by the twin constraints of fuel shortages and increased carbon emissions. Green, environment-friendly, and omnipotent photovoltaic (PV) solar and wind energy systems are sustainable alternatives to conventional energy systems. Moreover, Indian government policies propitious to the deployment of PV systems and wind systems have reduced the cost of electricity production. In India, consumption of energy has increased massively and has led to the development of renewable energy sources as an alternative to the fast-depleting non-renewable energy sources [1]. Utilization of resources such as solar, wind, fuel cells, and biomass has paved the way for technological advancements in the modern electric power sector. A smart grid is a digital technology that helps minimize or prevent power quality issues by integrating multiple microgrids with the grid and monitoring the microgrids and grid with proper management and control. Interconnected microgrids bolster the likelihood of compliance with the stability requirements of individual microgrids. Therefore, to reduce carbon emissions and provide an energy management solution, self-sustaining smart grid technology is required [2] Remedies for improving the quality of power in the grid include the use of passive or active filters. The bulkiness and performance issues of passive filters have led to a preference for active filters, despite their higher cost [3]. Since the advent of the present millennium, deployments of utility grid-interfaced PV and wind energy systems as active filters have been flourishing as an important area of research interest across the world [4]. Ample literature has been created to improve the active filtering capabilities of smart grid systems that are integrated with microgrids. Various algorithms used for controlling shunt active filters (SAFs) [5] include instantaneous reactive power theory (IRPT) [6], synchronous reference frame theory (SRFT) [7], Icos$\varphi$ [8], and improved linear sinusoidal tracer (ILST) [9]. These algorithms, when used for controlling the voltage source inverter (VSI) in grid-connected systems, can eliminate current harmonics on the source side and provide reactive power compensation [10].

1.1. Smart Microgrid

India’s Model Smart Grid Regulations [11] define a “smart microgrid” as an intelligent electricity distribution system that interconnects loads, distributed energy resources, and storage within clearly defined electrical boundaries to act as a single controllable entity with respect to the main grid [12]. In 2017, according to the International Renewable Energy Agency (IRENA), there was about 2179 GW of renewable energy source (RES) capacity. Wind energy was estimated at 514 GW, while solar energy was estimated at 397 GW. Approximately 76 GW of India’s 350 GW installed generation capacity (as of January 2019) was derived from RESs. India is proceeding toward reaching parity in power production with the estimated aggregated demand of the population, joining the ranks of other advanced economies such as USA, UK, Brazil, Japan, Germany, and France. The installed capacity of RES rose to 72 GW in 2018, and is expected to reach 175 GW by 2022 [13].

Microgrids (MGs) are small-scale power stations with power ranging from 100 kW that can operate in grid-connected and islanded modes, with high energy security, reliability, storage size, and economics for demand-side and load-side management [14]. They are capable of local control via automated mechanisms, and by inspecting and responding to problems, self-recovery of the power sector is possible. India’s electric sector, with solar and wind RESs, in combination with smart grid systems, has been transformed to return excess energy back to local microgrids and thus generate cost-savings. Smart microgrids are self-sufficient and can serve a local community without relying on central power grids, but they also have the ability to integrate with the grid. In [15], grid-tied solar PV is explored using a Kalman filter-based controller that minimizes the harmonics. The management of a hybrid photovoltaic system and battery storage for power flow control is examined with an artificial neural network in [16]. The growth of the smart grid has increased the opportunities for improving the demand response of consumers. In [17], the intelligent techniques for optimum scheduling of electrical energy is considered. Based on similar investigations made on electrical systems, intelligent controller-based smart grid systems with multiple MGs integrating with the grid without disturbing the quality of power will be a necessity of future generation.

Figure 1 shows a smart microgrid with solar PV and wind as RESs. PV systems have two types of configurations—on-grid and off-grid systems [18]—including technology constituted of energy in centralized and distributed modes. The off-grid system requires charged controllers for charging and discharging batteries, and inverters for converting the solar PV-generated direct current (DC) to alternating current (AC) at the consumer side. In an on-grid PV system, the output power of the PV is fed directly into the grid, which does not require a battery for storage. Wind energy systems can be used with other energy sources for better power management due to the intermittent nature of RESs. Wind systems can have predictable uncertainties based on location, time of day, the charge status of the battery, load variations, etc.; besides, there are unpredictable uncertainties such as weather conditions. Some of the predictable uncertainties are considered along with the solar PV source and wind energy source. This paper reviews the management of active power flow to sustain reliable and stable operations of the central grid.

1.2. Smart Microgrid Systems

A smart microgrid system is a collection of multiple smart microgrids linked together by an efficient controller, which can be integrated with the grid or operate independently, as depicted in Figure 2. The smart microgrid system comprises various renewable and non-renewable energy sources, as well as residential, commercial, and industrial load centers. Real-time management of electric power systems calls for statistical analysis of the collected consumers’ energy usage data, along with weather forecasts for different RESs using data analytics and artificial intelligence (AI) techniques. The AI-based controller processes information from the power consumers, together with other modules such as tariff [19] control and power flow management.

This paper focuses on the development of a smart microgrid system with multiple RESs integrated with the electrical power grid. The SAF control mechanism conceived here is based on the Icos$\varphi$ algorithm. This algorithm was upgraded using a fuzzy logic controller (FLC) and rendered intelligent by accounting for diverse parameters [20] regarding the main power grid, microgrid, source current, and load demand. This concept is extensible to any number of microgrids for the management of power flow targeted to enhance power quality. A detailed analysis of the MATLAB simulation and hardware model of the proposed FLC-based intelligent controller for active power-sharing and reduction of power quality problems under unbalanced nonlinear load are discussed in this research work.

The rest of the contents of this paper are organized as follows. Section 2 explains the materials and methods of the proposed smart grid system with multiple smart microgrids coupled with RES. Section 3 covers results, with design verification and testing of proposed smart microgrid system using intelligent integrated FLC in simulations and hardware platforms. Section 4 gives a brief discussion of the performance of the proposed smart microgrid system with the intelligent integrated controller (IIC) and its importance in the emerging research field. The economic feasibility of the smart microgrid model is also discussed in detail. Section 5 presents closing remarks, including the need for additional investigations.

2. Materials and Methods

2.1. Proposed Smart Grid System with Multiple Smart Microgrids Coupled with RES

The proposed smart microgrid system with the IIC is shown in Figure 3. The smart microgrid system comprises two microgrids—Microgrid 1 and Microgrid 2—integrated with the main grid. Microgrid 1 is powered by a PV panel and Microgrid 2 is powered by a wind energy source that is connected to the inverter for integration with the AC grid. Thus, the microgrids coupled to the main power grid act as SAFs to reduce the current harmonics due to the presence of nonlinear load, and improve the quality of power by sharing power during times of high demand. SAF DC-link voltages [21] and the microgrid output voltages must be the same for proper integration. This is possible by designing a boost converter that helps to integrate the two microgrids to participate in power-sharing. Microgrids interconnected with the grid at the point of common coupling (PCC) ensure reactive and active power compensation. SAF based on VSI is used to connect various microgrids to the grid, as well as to eliminate current harmonics and compensate for reactive power.

Figure 3a shows an electrical network of the smart microgrid system with the power grid integrated with two microgrids combined with transformers, circuit breakers, and various sources, including renewable and nonrenewable sources feeding different loads. The smart microgrid system comprises two microgrids—Microgrid 1 and Microgrid 2—integrated with the power grid. Figure 3b,c show the detailed analysis of Microgrid 1 and Microgrid 2 with RES mentioned in Figure 3a. Microgrid 1 has a solar PV array as RES and Microgrid 2 has a wind turbine generator as the source. The battery bank with a bidirectional DC–DC converter connected to the DC link is used to ensure charging and discharging. With the help of the SAF-based VSI, the microgrids are integrated with the PCC of the electric grid. The maximum power point tracking (MPPT) and charge controller ensures maximum power tracking of the RES and the charge control of the energy storage system [22]. Different industrial, residential, and commercial consumers with linear and nonlinear devices are linked to the AC link voltage through smart meters. These nonlinear devices create disturbances in the quality of power and hence, the electric grid becomes disturbed by distorting the source voltage and source current, thereby increasing the harmonics [23]. The proposed controller was used to tackle all such problems by ensuring active power support with harmonic elimination and reactive power compensation.

2.2. Proposed AI-Based Icos$\varphi$ Control Algorithm for Power Sharing and Power Quality Improvement in Smart Microgrid System

The RES-powered microgrids integrated with the main grid in the presence of nonlinear loads produce harmonic currents that are in phase opposition to the reactive currents present in the load. The power quality problems created by non-linear devices can be eliminated by connecting SAFs, as shown in Figure 3a–c. Here the IIC uses an efficient active filter control algorithm to provide the compensation current to be injected by the SAF. They produce harmonic currents that are in phase opposition to the reactive currents present in the load. An efficient active filter control algorithm is required to provide the compensation current to be injected by the SAF. Here, the control algorithm used is the Icos$\varphi$ algorithm, modified using the intelligent FLC, which is explained in detail.

The balanced three-phase source voltages are

$$v_a=V_{m}sin\left(\omega t\right)$$

(1)

$$v_b=V_{m}sin\left(\omega t-120\mathrm{°}\right)$$

(2)

$$v_c=V_{m}sin\left(\omega t+120\mathrm{°}\right)$$

(3)

An unbalanced nonlinear load is connected and the reference source current is calculated by taking the average of the magnitudes of the real components of the fundamental load currents of the three phases, as shown in Equation (4).

$$I_{sref}=\left(\mid I_{La}\mid cos\left({\varphi }_a\right)+\mid I_{Lb}\mid cos\left({\varphi }_b\right)+\mid I_{Lc}\mid cos\left({\varphi }_c\right)\right)÷3$$

(4)

The compensation current provided by the SAFs for phases a–c is obtained as in Equations (5)–(7), respectively.

$$i_{fa}=i_{La}-i_{rsa\left(ref\right)}$$

(5)

$$i_{fb}=i_{Lb}-i_{rsb\left(ref\right)}$$

(6)

$$i_{fc}=i_{Lc}-i_{rsc\left(ref\right)}$$

(7)

where i_La, i_Lb and i_Lc are the actual load currents; i_rsa(ref), i_rsb(ref), and i_rsc(ref) are the reference source current for the a–c phases respectively.

In this paper, an intelligent control algorithm is proposed as shown in Figure 4 by making modifications to the Icos$\varphi$ algorithm with gain factor ‘K’, the output of the intelligent fuzzy controller. This controller takes input parameters such as state of charge (SoC) of microgrids, source current, and the tariff of microgrids, and derives the amount of power that needs to be shared by the grid and microgrid. The ‘K’ factor is multiplied by the reference source current and then subtracted from the load current to calculate the filter current in each phase a–c, as in Equations (8)–(10) respectively, which reduces the harmonics by providing reactive power compensation.

$$i_{fa}=i_{La}-K{i}_{rsa\left(ref\right)}$$

(8)

$$i_{fb}=i_{Lb}-K{i}_{rsb\left(ref\right)}$$

(9)

$$i_{fc}=i_{Lc}-K{i}_{rsc\left(ref\right)}$$

(10)

The detailed analysis of the integrated intelligent controller is explained in Section 2.3.

2.3. Intelligent Integrated Controller for Smart Microgrid Using FLC

An intelligent controller can make decisions on constraints in real-time, conduct numerical computations with efficient processing, and sustain bidirectional communications with the smart microgrid system. The controller can determine the appropriate tariff rate for utility consumers regarding their daily energy usage profile. Based on energy drifts in the preceding years, the controller anticipates future energy demand. The controller can decide which supplier is best qualified to participate in power-sharing modulated by the microgrid tariff rates and available power.

The FLC is a decision-making tool that allows for the definition of intermediate values between true or false. Identifying a fuzzy system involves two phases: identifying the structure and predicting the parameters. In terms of structure identification, determining input variables, selecting the fuzzy rule system, and determining the number of fuzzy rules and type of membership functions is required [24]. The importance of FLC techniques with triangular membership functions in the field of engineering for prediction is analyzed in [25]. Analysis of developing a single-phase series active filter using Mamdani’s model FLC, sliding mode controller, and PI controller is presented in [26]. Triangular membership functions can be used to determine gain factor, error, and derivative of error for better controller action. Sharing power among storage devices in an interconnected microgrid system with an FLC-based system with triangular membership function is implemented in [27]. The fuzzy logic energy management for residential systems with grid connection using triangular function is discussed in [28]. The literature elucidates the effectiveness of triangular membership function in this field. Depending on the range and variability of the input and output data, Mamdani model FLC with a triangular membership function is found to be effective and is accordingly used in the proposed system.

The proposed FLC architecture for a smart microgrid system is shown in Figure 5, with inputs including the source current, the SoC of the microgrids, and both microgrid tariffs. The two microgrids with RESs integrated with the three-phase ac grid are connected with an unbalanced diode bridge rectifier load. As the load connected is nonlinear, there will be distortions in both the source and the load. The inputs of the FLC can be source current (I source), SoC, and tariff of microgrids.

The SoC plays a critical role in determining how efficiently the controller will handle changes in the operating conditions of the smart microgrid system.

Table 1. Range of source current.

Source Current	Range
I1	0–5–10
I2	9–12–15
I3	14–18–22

,

Table 2. Range of state of charge (SoC).

SoC	Range
S1 (very low)	10–25–40
S2 (low)	35–50–65
S3 (medium)	50–65–80
S4 (high)	70–85–100

and

Table 3. Range of tariff.

Tariff	Range
T1	0–0.2–0.4
T2	0.3–0.5–0.7

shows the range of source current, SoC, and tariff, respectively. The gain factor ‘K’ shown in

Table 4. Ranges of gain factor (k).

Gain Factor (K)	Range
k1	0–0.35–0.7
k2	0.6–1–1.4
k3	1.3–1.5–1.7

is the output of the controller where the range is selected following different operational conditions. k1 is the gain factor ranges where the microgrid enters the mode where the stored energy is sold to the grid. At k2, the microgrid enters a mode where the active, reactive, and harmonic support will be provided by the microgrid, and at k3 the microgrid enters a mode where compensation and charging occur side by side.

The fuzzy membership function is used to fuzzify the categorized sets. As the variation across the classified ranges is linear, we employ triangular membership functions for the fuzzification. To ensure the system’s ability to reduce the calculation error at the point of contact of membership functions, fuzzy sets are examined in such a way that membership functions overlap. The membership function plots in Figure 6, Figure 7 and Figure 8 show a condition where the source current is low and the SoC of the microgrid is very low, with a low microgrid tariff condition. The gain factor K is obtained as k1, as in Figure 9, indicating that the renewable energy stored in the microgrid is given to the grid.

The rule viewer of the FLC for medium source current, a very low value of SoC of the microgrid, and high tariff for the grid is illustrated in Figure 10. It is observed that K is 1, which implies k2, as seen in Table 4. This signifies that the microgrid participates in active power support and in providing compensating signals. From Figure 11, when the source current is medium, the SoC of the microgrid is very low and the grid tariff is low; K = 1.5 implies k3. In this mode, the microgrid enters charging mode where both compensation and charging occur side by side. A three-dimensional view of k1, k2, and k3 is shown in Figure 12.

3. Results

3.1. Design Verification of Proposed Smart Microgrid System Using Intelligent Integrated FLC

Microgrids integrated into the grid with a three-phase nonlinear load were modeled and tested by MATLAB simulation. The proposed intelligent controller with modified Icos$\varphi$ as the filtering algorithm was used for the microgrids to operate with an SAF. Two microgrids are integrated separately first and Fast Fourier transform (FFT) analysis was performed. Later both the microgrids—Microgrid 1 and Microgrid 2—were integrated together and the results were analyzed and discussed. The simulation model of the proposed system with unbalanced nonlinear load is analyzed and discussed in Section 3.1.1. The performance study of the smart microgrid system with the intelligent integrated FLC, which incorporates tariff and power flow management and can lessen the stress on the main grid, is explained using a MATLAB simulation modeling in Section 3.2. The hardware implementation was done in the following stages: load selection and testing, design and open loop testing of the control algorithm, design of sensor circuits, pulse generation and driver circuit, open loop testing of the IIC, and testing of the IIC with sensed parameters from power circuit, which is explained in Section 3.3.

3.1.1. Simulation Model of the Proposed System with Unbalanced Non-Linear Load

The proposed smart microgrid was tested using a MATLAB Simulink model with a grid of 415 V, 50 Hz three-phase source connected to a diode bridge rectifier load. The microgrids taking part in power-sharing have a DC output voltage of 650 V. The three phases of the system are connected to a diode bridge rectifier load and are unbalanced, as shown in Figure 13.

The simulation of the smart microgrid system was carried out for 2 s. At 0.2 s, a load of 5 kW was added, and at 0.5 s, another 5-kW load was added. As a result of the increase in the load in each phase, the load current was also varied accordingly. The state of charge of the microgrid with the PV source was varied at 0.1 s, 0.3 s, and 0.4 s, resulting in a decrease in source current. The source voltage, source current, load current, and compensating current generated by Microgrid 1 at phase ‘a’ when the PV microgrid is integrated into the grid are shown in Figure 14.

The total harmonic distortion (THD) of the grid before compensation and after compensation is displayed in Figure 15a,b, respectively. The THD value before compensation was found to be 30.66%, and after compensation was found to be reduced to 6.29%.

As the next step, Microgrid 2 with a wind RES was interfaced with the main grid, with a higher SoC value. The performance of the IIC with Microgrid 2 when connected to an unbalanced nonlinear diode bridge rectifier load was analyzed. In this case, the smart microgrid system with IIC was also simulated for 2 s; the load was varied at 0.2 s and 0.5 s and the SoC level was varied at 0.1 s, 0.3 s, and 0.4 s. Thus, the source current and the compensating current vary according to the changes, as in Figure 16. Thus, when both the microgrids are integrated into the main grid, the FFT analysis after compensation is shown in Figure 17. Since the SoC level of Microgrid 2 was higher than Microgrid 1, the THD value after compensation was found to reduce to 5.58%.

3.2. Tariff Controller and Power Flow Management of the Smart Microgrid System

The tariff controller checks the tariff of both microgrids and the load demand and the SoC of the microgrids, and performs such that the consumer will choose the microgrid with a minimum tariff based on the load demand and available power. While considering two microgrids, the tariff of each microgrid was compared at each instant of time, along with the other two parameters: the SoC and the source current. The control signal for the solar PV microgrid and wind microgrid generated in Figure 18 was such that the Microgrid 1 tariff as compared to the Microgrid 2 tariff was higher up to 0.1 s, lower up to 0.4 s, and higher after 0.4 s. If both the microgrids operate with the same tariffs, both will give the same compensating signals to the grid. Figure 19 shows voltage and current at each microgrid based on tariff. The voltage and current of the grid become perfectly sinusoidal and distortion-free. The load was varied at 0.2 s and 0.5 s as before. The source current was found to vary slightly due to the changes in SoC levels.

The power flow management under different tariff conditions and equal tariff conditions are shown in Figure 20a,b, respectively. It was inferred that both the microgrids shared power based on tariff controller and intelligent FLC. From 0 to 0.1 s, the gain factor of FLC was k1 and the tariff of Microgrid 2 was lower; hence, Microgrid 2 and the grid provided the active power required for the load. From 0.1 s to 0.3 s, the gain factor of Microgrid 1 was k2, Microgrid 2 was k1, and the tariff of Microgrid 1 was lower, so Microgrid 1 and the grid provided the active power required for the load. From 0.3 s to 0.4 s, the k factor of Microgrid 1 and Microgrid 2 was k1, and the tariff of Microgrid 1 was lower; hence Microgrid 1 and the grid provided the active power needed for the load. After 0.4 s, the gain factor of Microgrid 1 was k2, Microgrid 2 was k1, and the tariff was lower for Microgrid 2; hence, Microgrid 2 and the grid provided the active power needed for the load.

From the observations, it is seen that the grid, Microgrid 1, and Microgrid 2 share power equally with the load. The THD of a compensated system under different tariff conditions is 4.92%, as in Figure 21a, and at an equal tariff was reduced to 3.64%, as in Figure 21b.

Table 5. THD of smart microgrid system for different conditions.

System	THD
THD of uncompensated system	30.66%
THD of smart microgrid system with MG1 integrated to the grid with IIC.	6.29%
THD of smart microgrid system with MG1 and MG2 integrated to the grid with IIC.	5.59%
THD of smart microgrid system with MG1 and MG2 integrated to the grid with IIC at different tariffs.	4.92%
THD of smart microgrid system with MG1 and MG2 integrated to the grid with IIC at equal tariffs.	3.64%

gives insight into the THD of the smart microgrid system under different conditions, suggesting that the proposed integrated intelligent controller works efficiently for reducing the harmonics and improving the power quality of the system with tariff control and power flow management. The performance of the proposed intelligent integrated controller provides active power support with improved power quality when both microgrids have equal tariff conditions.

3.3. Hardware Implementation of the Proposed Intelligent Integrated FLC

The testing of the system in hardware was done with a prototype model, as shown in Figure 22. The three-phase ac supply of 415 V used in the simulation was scaled down to 50 V using a three-phase autotransformer. A 50W, 3$\varphi$ diode bridge rectifier feeding resistive loads was chosen as a harmonic load to be connected at the coupling point. The harmonic analysis of the load was done using a power quality analyzer by connecting the load to a 50 V, 3$\varphi$ source. The intelligent integrated FLC-based Icos$\varphi$ controller was built using an ARDUINO microcontroller, which generates pulses for grid integration based on the sensed load currents and grid voltages. An AC of 1.2 A was found to flow through the circuit. The current across the diode bridge rectifier was found to be 1.5 A. The reference filter current obtained from the Icos$\varphi$ was compared with the ground, and using a TLP driver circuit, the pulses for the inverter was boosted to 15 V. On the AC side of the inverter, an RL load was connected and using this, inductor tuning of filter current was performed.

The reference source currents obtained from the Icos$\varphi$ algorithm were compared with the actual values to generate pulses for switching the inverter. To compare the reference and actual values, a LM324 comparator IC was used. The pulses for the lower arm of the inverter were obtained by complimenting the upper arm pulses using IC 7404. The pulse output of LM324 and 7404 was in the range of 5V. TLP250 was used for boosting the output voltage and to provide isolation. Figure 23a,b shows the output pulses from TLP given as input pulses for switching the inverter’s first, third, and fifth thyristors and second, fourth, and sixth thyristors, respectively.

Figure 24a–c shows k values, reference source current, and k* irs(ref) current waveforms at different k gain factors. From the proposed AI-based Icos$\varphi$ controller, when the current was 1 A, SoC was 70%, and when the tariff was high, the k factor was found to be 0.35. Similarly, when the current was 1 A, SoC was 50%, and when the tariff was high, we obtained k as 0.85. The value of k was found to be 1.34 when the current was 1 A, SoC was 25%, and the tariff was low.

From the equations of the proposed AI-based Icos$\varphi$ controller for power sharing and power quality improvement discussed in Section 2.2, we know that the filter current generated must be the difference between the actual load current and the product of the gain factor and reference source current, as shown in Equations (8)–(10) for the a–c phases, respectively. Observing the waveforms of Figure 25a–c, we can infer that the load current (IL), k* irs(ref) current, and filter current obtained for k at 0.35, 0.85, and 1.34, respectively, satisfy the requirements of our proposed controller.

4. Discussion

The performance of an IIC with SAF in the Simulink model was analyzed and the controller was found to work effectively. The smart microgrid system with the controller was validated with different values of gain factors k—k1, k2, and k3—based on different input parameters such as the SoC of the battery, load demand, and the tariff of the microgrids in the simulation platform. The results were found to be accurate based on the THD analysis and the power flow management with tariff control. The performance analysis of the tariff controller with the power flow management unit was discussed in detail and it was found that proper active power-sharing is achieved along with reactive power compensation, as explained in Section 3.2 in Figure 21a,b. FFT analysis validated the efficient working of the proposed smart microgrid system. The THD, which was 30.66% for the uncompensated system, was reduced to 3.64% under equal tariff conditions. A prototype laboratory model of the intelligent controller with a three-phase diode bridge rectifier load was implemented. The hardware results highlight the performance of the IIC under low, medium, and high conditions of source current, SoC, and tariff.

Economic Feasibility of the Proposed Smart Microgrid Model

The economic viability of the smart microgrid system is an important factor of concern. The proposed smart microgrid system is multiple microgrids integrated to the grid with tariff control, ensuring proper power flow between microgrids and the grid by maintaining the quality of power. The cost–benefit analysis (CBA) is one of the major methods through which economic aspects are dealt with in detail [29]. For CBA, the benefits of the proposed system are to be taken into consideration in the first step. In the next step, the cost of the system has to be identified and as a final step, both the cost and benefits must be compared to obtain the economic viability of a system. For finding the CBA of the system, the baseline is a grid-tied system without distributed generation (DG) capable of feeding a small residential area with around 20 houses. The energy consumption of each house per month can be approx. 300 units, and hence, for one year (12 months), the consumption will be 3600 units. For 20 such houses, the total units consumed is 72,000 units per annum and the annual electricity cost will be Rs 7.2 lakhs at Rs10/unit. Considering the proposed model with multiple microgrids integrated with the grid, the initial investment includes the cost of various equipment such as multiple microgrids with PV and wind of 10kW each, the IIC, smart inverters, and converter circuits, smart sensors, smart meters, power quality analyzer, etc., ranging to Rs 25 lakhs. Microgrid 1 and Microgrid 2 include RESs with solar panels and a wind source of 10 KW capacity each, capable of generating 50–60 units/day. Each of the MGs will generate 1800 units/month and hence 72,000 units per annum. In the model proposed here, we have seen that when both MGs have an equal tariff, the power flow is such that only 1/3rd of the load is met by the grid and the remaining 2/3rds are shared by both MGs. Hence, the units consumed by the residential load of 20 houses is 1/3rd of 72,000 units, equal to 24,000 units. which is shared by the grid at Rs10/unit energy consumption, with the remaining 48,000 units shared by both MGs at Rs 4/unit energy consumption.

$$Totalcostflowperyear=\left(24000\times 10\right)+\left(48000\times 4\right)=Rs432000=Rs4.32lakhs$$

(11)

$$AnnualSavings=\left(Rs7.2lakhs-Rs4.32lakhs\right)=Rs2.88lakhs$$

(12)

$$PaybackPeriod=InitialCost÷AnnualSavings=Rs25lakhs÷Rs2.88lakhs=8.7years$$

(13)

5. Conclusions

This paper presented a smart microgrid system integrating multiple microgrids with RES using an AI-based Icos$\varphi$ controller for power sharing and power quality improvement. The integration of two microgrids with RESs is a very fast growing area, since RESs are intermittent in nature and also reduce carbon emission in the environment. However, as we discussed in the introduction section, there are many factors to be taken into account while developing the proposed system, among which power sharing and power quality are given focused on here. Two different RESs, such as solar for Microgrid 1 and wind for Microgrid 2 are integrated with the grid, which has an unbalanced three-phase diode bridge rectifier load. The challenges faced during the process include designing an IIC with a proper control algorithm that can make the SAF to reduce the harmonics within the IEEE standards and provide reactive power compensation with proper energy management. An AI-based IIC is implemented in this research, which is capable of providing energy management by taking into account various parameters such as the SoC of the battery of the MGs, source current, and the tariff of the MGs. The THD of the smart microgrid system at different conditions in Table 5 shows the efficiency of the proposed AI-based IIC. The design of the proposed AI-based IIC was tested and proven with the hardware prototype for all conditions of the K factor, and the reference source current, load current, and the filter current obtained in Section 3.3 affirm the productive function of the IIC, which can be used for further research developments. The economic feasibility of the system was investigated, and it was observed that the system has a payback period of 8.7 years, which is adequate for a renewable energy integrated system. As a future scope, this research enhances the intelligence of the proposed AI-based IIC by forecasting the output power of the microgrids with RESs, thus achieving power flow management with tariff control without affecting the quality of power. Administration of the IIC in high voltage system with real time hardware in loop simulation platform is also a very relevant topic for future investigation.