Residential Community Load Management Based on Optimal Design of Standalone HRES With Model Predictive Control

Microgrids being an important entity in the distribution system, and to get their full advantages by incorporating maximum distributed generation, standalone hybrid renewable energy systems (HRESs), being environmentally-safe and economically-efficient, are considered as the promising solution to electrify remote areas where the grid power is not available. In this work, a techno-economic investigation with an optimal design of HRES is presented to fulfill the domestic electricity need for a residential area of the Sherani district in the Province of Baluchistan, Pakistan. Nine case studies based on PV/wind/diesel/battery are analyzed based on net present cost (NPC), cost of energy (COE), and emission to decide the feasible solution. HOMER tool is utilized to accomplish modeling and simulation for economic analysis and optimal sizing. Simulation results demonstrated that HRES with PV-wind-battery is the most viable option for the specified area, and the optimal sizing of components are also obtained with $ 28,620 NPC and $ 0.311/kWh COE which shows 81.65 % reduction in cost and 100 % preserving in toxic emission while fulfilling 100 % energy demand with 67.3 % of excess energy. Furthermore, MATLAB/Simulink modeling for the optimally designed system is built for technical analysis while its effectiveness is proved by keeping dc and ac buses voltage constant, safe operating range of battery state of charge (SOC) with active power balance between HRES components, as well as efficient ac voltage quality, regardless of generation disturbances and load fluctuations. The output signal has total harmonic distortion (THD) of 0.30 % as compared to 5.44 % with the conventional control scheme. The novelty lies in the sequential application of both HOMER and MATLAB simulations of the proposed HRES model and validation of the proposition for the studied area; by using and implementing model predictive control (MPC) of a reconfigurable inverter.

Mechanical power from the wind VOLUME 8, 2020

I. INTRODUCTION
Globalization and inter-countries energy sharing are considered as the conceivable future for the globe. In the present time, a significant proportion of the world's population is still living in remote rural areas with almost no access to electricity. Around 17 % of the world population (1.2 billion) cannot use electricity [1] and almost 80 % are living in rural spots [2]. Meanwhile, about 2 billion population around the globe have no access to grid-based electricity according to the most recent statistics [3]. Mostly, these peoples live in remote areas with no access to the national grid. According to [1], biomass is being used by around 3 billion worldwide habitats for cooking, heating and lighting purpose and resulted in 3.1 million premature deaths due to air pollution with incomplete biomass combustion. This is highlighted as the main reason for cramping the economic and social development of such communities in developing countries like Pakistan. Moreover, greenhouse gas (GHG) emissions are produced from many sources as shown in Fig. 1 in which it is noted that the dominant sector of producing these emissions is the energy sector through the conventional ways of energy production, while residential buildings indirectly produce the largest proportion of CO2 [4].
Conventional energy sources (CESs) cannot meet the day after day growth in energy demand due to their expensive capital cost, limitations of its fuel and the infeasibility of grid extensions due to geographical locations. Furthermore, as mentioned, CESs are the major contributor to greenhouse gas (GHG) emissions and the most severe threat to the environment and human health. Due to what has been mentioned, establishing small-scale grids (microgrids) is considered a shining solution to face most of the above-mentioned obstacles. Renewable energy resources (RERs) are the more feasible option to tackle all modern global warming and depleting fuel issues. These RERs (e.g. PV, wind turbines, fuel cells) can be easily integrated with each other in different microgrid (MG) configurations for electrification of remote and urban areas. The various hybrid configurations of RERs are adopted in literature and are considered as the most cost-effective, and environmentally friendly. However, its intermittent nature is considered as the main drawback which needs to be handled carefully with efficient energy management and control (EMCS) algorithms. Among the different control techniques that can be used to achieve the EMCS, Model Predictive Control (MPC) is utilized because of the reason that this algorithm is mainly classified as one of the most powerful artificial intelligence and non-linear programming approaches.
Due to the merits of MPC in addressing the systematic processing of constrained multivariable, it is well developed to handle the issues of power systems and power plants [6]. Hierarchical and distributed MPC is more efficient for handling large and complex power system problems through future prediction of control actions. Finite control set MPC (FCS-MPC) is the most frequently used among MPC controls due to its simplicity and high accuracy [7]. MPC is used in various applications including load frequency control for dynamic performance improvement under various disturbances and many applications in electrical drives (speed/position/torque control, torque ripple reduction, fieldoriented control) [8]. VSI with LC filters is widely used in power supplies for high-quality output, which is also the future application in an embedded network of electrical aircraft [9].
Currently, renewable energy share worldwide is only 11 % while it is expected to increase by 60 % in 2070. The global capacity of wind and solar photovoltaic (PV) is increased to 514.8 GW and 399.6 GW respectively [6]. Fig. 2 shows the global trend for investment in the renewable energy sector. About the GCC countries [10], RERs penetration into the existing grid is an economically viable option for diversifying electricity mix and creating an avenue. Multi-directional power flow with a wide distribution system is the possible future of power grids in GCC states. Solar energy is going to be more competitive in all GCC regions and hence private firms would play their role in this regard. Proper deployment of new technologies (i.e. smart grid and smart cities) would be the economic growth factor with more revenue as well as the political stability of GCC countries with neighbors. The most feasible GCC regions for technology implementation are UAE, KSA, and Kuwait [10].
Pakistan is facing a severe problem of energy shortage and this shortfall is increasing with each passing day. Over 51 million peoples in Pakistan are still living with off-grid access to electricity [11]. Pakistan has initiated different projects in all provinces to curb the energy shortage problem. Potential areas for wind power projects include four sites in Punjab, twenty areas in Sindh, 25 points in Khyber Pakhtunkhwa, 23 sites in Baluchistan, two in FATA, three in Azad Jammu and Kashmir, 11 in Gilgit Baltistan [12]. World Bank has installed MHP weather stations in nine cities with consideration of five family members in one household [13]. In Pakistan, key points of national power policy 2013 include load shedding elimination, decreasing transmission and distribution losses from 23-25 % to 16 %, increasing revenue from 85 % to 95 % and time reduction for decision making. Power policy 2015 focused on private FIGURE 2. Global trends of renewables [16]. investment in the energy sector for sufficient power generation capacity, least COE, priority of indigenous resources' use, facilitating all stakeholders and safeguarding the environment [14].
Pakistan is bestowed with a high potential for RERs (solar and wind). These RERs are not comprehensively explored so far. The country is dependent mainly on thermal, hydro and nuclear energy, which creates acute energy shortage problems nationwide especially in rural areas where grid power is not available. Remote area electrification is normally carried out by expensive grid extension, small hydro and diesel power plants, which have a severe impact on health as well as environment. It is expected to achieve a target of 5 % renewable energy mix until 2030, currently, there is no significant penetration of renewable energy. Renewable energy activity in Pakistan is considered very promising to be exploited, the daily solar radiation ranges from 4.68 kWh/m 2 /day to 5.54 kWh/m 2 /day while the annual global irradiance range lies between 1900-2200 kWh/m 2 . Additionally, about 346 GW wind potential is also available. As this study aims at the electrification of one of the Pakistani rural communities, one household maximum load is considered to be 21.57 kWh which is also shown in Table 5 while energy ranges for domestic community load are shown in Table 1 [15].  [15].
The organization of the remaining paper is presented as follows; a literature summary and comprehensive comparative analysis are established and described in section II along with the main objectives, scope, and significance of the proposed scheme. Section III presents the detailed system description which includes feasibility study (site/resources selection, load estimation), and modeling of HRES components (PV, WT, DG, battery, and power converter). Detailed discussion on techno-economic analysis using HOMER is planned in section IV which includes the optimal design objective and constraints supported with a detailed flow chart. Section V investigates the technical aspects using converters controllers in MATLAB/Simulink for the most optimal plan obtained from section IV while section VI describes the simulation results with detailed discussion for the suggested model. Section VII compares and verifies the key results of the proposed model with literature studies along with the main findings. Finally, the conclusion is grasped in section VIII.

II. RESEARCH BACKGROUND SYNOPSIS
Several researchers emphasized the study of standalone HRES. Table 2. summarizes the addressed researches that investigated such kind of systems, such as the configuration of the HRES, the applied algorithms as well as the main objectives are presented. Table 3 shows a comprehensive study of the research methodologies of the literature in terms of their contribution and drawbacks. According to the performance rating for the suggested energy management and control of HRES, a comparison between suggested work and presented literature is shown in Table 4. Results show that most of the literature investigated PI-based control but thorough investigation on the basis of design optimization is not presented. The wind control unit in the current study includes an uncontrolled rectifier which is simple, and economical. Moreover, an uncontrolled rectifier works without IGBTs and there is no need for additional controllers for control signals and thereby reducing the complexity and system cost. Most of the published literature was not thoroughly analyzed by considering wind and PV MPPT, uncontrolled wind rectifier, battery controller for regulating dc voltage, THD analysis, comparison of PI and FCS-MPVC for inverter control.
Most of the addressed literature studies have been solely concentrated on either design optimization of HRES or energy management system (EMS). In this work, the optimal design of HRES with the most suitable and advanced EMS approach is proposed. First, the detailed technoeconomic analysis with multiple hybrid scenarios for diesel, solar photovoltaic (PV), wind, battery, and converter are    applied for the design optimization of HRES. The developed model is then utilized to fulfill the load demand of an islanded domestic area sited in Pakistan with the consideration of real demand profile and weather information. Second, MATLAB/Simulink R is used to develop a self-made simulation tool for the implementation of the suggested EMS. The optimized model is tested on the proposed EMS and the performance of the suggested model is improved for power quality, steady-state and transient stability with intermittent generation and fluctuating load demand.
Summarized points from the comprehensive literature review are: • The HRES configurations are considered to be the most reliable and economically viable options with minimum GHG emissions. While multiple configurations are possible for any specific location based on load profile, resources data, and environmental conditions.
• The major focus of the literature is either on optimal sizing with economically viable configuration or on power management of HRES. Both features are not given much attention at the same time. According to the best of authors' understanding, the proposed strategy by handling optimal design and power management simultaneously for the PV-WT-Battery system using FCS-MPC with reconfigurable inverter for a standalone HRES is not investigated in the literature.
Based on the above-mentioned circumstances, the objectives and scope of the presented work are summarized as follows: • To fulfill the drastic increase in electricity demand, the sample location of the Baluchistan province in Pakistan is investigated with the available energy resources including PV and wind.
• A unified and generic methodology for optimization of components size with power control and energy management for off-grid HRES is presented. The proposed methodology is endorsed with a real domestic case study for electrification in Pakistan.
• The optimal HRES scheme is obtained with the aid of feasibility study keeping in view the minimum cost and least emission. Minimum TNPC, LCOE, and improved model reliability are integrated as an objective function. HOMER R software is used for this purpose.
• Verification of a properly designed EMS is executed with the help of a self-made Simulink model in MATLAB R environment. The suggested EMS strategy is implemented with battery SOC to maintain load balance as well as dc bus voltage by keeping in view the battery SOC within the permissible limits, with maximum wind and PV power extraction while maintaining the ac bus voltage constant under fluctuating loads, and source disturbances (wind speed and solar irradiance).
MPC is the most preferable solution for complex control problems [55]. The advantages of the suggested FCS-MPC based model are listed as follows: • The significant advantage of MPC management is simplicity while the concepts are very heuristic and intuitive in nature and easily understandable.
• The suggested MPC based EMS model is more proper for multi-variable systems and can be extended accordingly.
• The suggested EMS model is applicable for real on-line systems of non-linear nature.
• The proposed FCS-MPC methodology with a complete solution for optimal design with energy management and control for the proposed domestic site is not analyzed in the literature.
• The extension of the system with the proposed FCS-MPC controller is very easy ranging from domestic to commercial and agriculture loads.
• The proposed control strategy with FCS-MPC selects the optimal strategy by controlling the interaction and VOLUME 8, 2020 constraints among different variables. Therefore, this strategy serves the purpose of economic benefits with the quick operation and predicts the dynamic behavior for linear as well as non-linear multivariable models.
• As compared to most widely used PI control, the suggested MPC based model can also be applied for gridconnected applications in addition to standalone systems during transient stability analysis.
• The settlement time for FCS-MPC control is less and is applied to minimize the errors and eliminate signal harmonics from the output voltage/current.
• The proposed control methodology does not need any modulation scheme and is applicable for variable switching frequency.
• The steady-state performance is better for all three reference frames with low design complexity and easy implementation of the resulting FCS-MPC controller in experimental studies.
• The current model based on PV-WT-DG-batteryconverter is the improvement of the WT-DG-batteryconverter model of [42] by incorporating a detailed analysis PV model.

III. SYSTEM DESCRIPTION
Pakistan is situated in South Asia with its 1046 km coastline alongside the Arabian Sea as well as the Gulf of Oman as shown in Fig. 3  The power sector of Pakistan is operated by WAPDA as well as KESC. KESC rebranded as K-Electric has the responsibility of generation, transmission, and distribution of the electricity to Karachi city and neighboring localities while WAPDA has the responsibility to provide generated power to the rest of the country. Under the new institutional setup, NEPRA is an independent regulatory authority for ensuring a transparent, competitive and commercial based power market. WAPDA is further divided into four public sector GENCOs, ten DISCOs, and NTDC. PPIB is established to facilitate the power sector with private investment and is handling the development of renewable energy resources (RERs). PAEC (Pakistan Atomic Energy Commission) develops nuclear power [14].

A. SITE DESCRIPTION AND ENERGY RESOURCE ASSESSMENT
The suggested methodology of HRES design and control is employed for the rural area electrification of a domestic load which is located in the province of Baluchistan Pakistan as shown in Fig 3 where the utility grid availability is obsolete. The wind speed range in Baluchistan province is observed to be 4 to 9 m/s and 12.5 m/s at 10 m and 50 m hub height respectively [59]. The designated location has Latitude and longitude of 31 •

B. ELECTRIC LOAD ESTIMATION
In this study, the local survey for the load estimation of one household is carried out in the Sherani district of Baluchistan. The energy consumption activities of prime importance are shown in Table 5. The estimated load demand is 3.62 kW (peak value) while the annual utilization is 21.57 kWh/day. Fig 9 shows the load profile for three typical houses of the selected location.  Fig. 10 shows the suggested strategy for the optimization of components size with power management for an offgrid HRES. The initial step involves the assessment of realtime meteorological information and the load profile of the suggested site. After specifying the optimization objectives along with constraints, feasible configurations of the HRES, the elaborated model of multiple model components is accomplished. The most feasible model is proposed based on the detailed analysis with the assessment of three aspects i.e. technical, economic and environmental. The next step involves the execution of the suggested EMS for the winning configuration plan. The final step analyzes the validation of the proposed management strategy and its comparison with the conventional method. The description of the proposed HRES, as well as the detailed modeling of system components, are discussed as follows:

C. HRES STRUCTURE AND COMPONENTS SPECIFICATIONS
The suggested system comprises DG, WT, PV, battery storage, and power converter. PV, Wind, and diesel are considered as the main generation to fulfill the energy requirements. The battery storage is considered to feed load under fluctuating PV and wind generation during different conditions (i.e. transients, ripples, and spikes). The power converters are an essential part of the HRES to supply energy from dc and ac bus. It is significant to mention that the boost converters are used for extracting maximum power from PV and wind, while battery converter (buck-boost) stabilizes the constant dc bus voltage.

1) WIND ENERGY CONVERSION SYSTEM (WECS)
The WECS contains wind turbines, PMSG, and the mechanism for maximum wind power extraction. Wind model from [42] of 1 kW rating is selected, while initial and replacement costs are set to $ 900/kW [61], O&M cost as $ 10/kW with 20 years project life. Wind mechanical power P mec is expressed as: The blade tip speed ratio is expressed as Fig . 11 shows the maximum values of the power coefficient (Cp) against the tip speed (λ). Normally, the value of the pitch angle is zero when P mec is below nominal power. Hence, Cp is the function of λ only and is maximum C pmax at a specified value of λ as shown in Figure. Blueline with zero angles (beta) shows that the tip speed of 8 units has a maximum power coefficient of 0.48 which is the case of the current wind model in this paper. At this moment, the WT operates at MPPT and the optimal rotor speed ω Dopt specific wind speed of v w in (4). By substitution of (4) in (1), we get: 3 .
The wind MPPT algorithm and the pitch angle control are implemented for maintaining active power (P). The reference P is obtained through the MPPT model based on the present rotor speed (ω D ). The relation for automatically reaching optimal rotor speed is expressed as follows: Keeping in view the fast-electronic operation, the wind power (P) is assumed to be equal to the reference power.

2) PV GENERATOR
The solar power is comparatively cheap as compared to wind power because of no wear and tear [6]. Incremental conductance (IC) is a commonly used method for solar PV MPPT which improves steady-state as well as dynamic response with good performance under fast-changing atmospheric conditions. Its basic principle for three MPP tracking conditions states that the power derivative is zero at MPP while it is negative on the right side of MPP and positive on the left side for power versus voltage curve. It can be expressed as [65].
The three MPP tracking conditions are represented by Eqs. (8)- (10). I / V is incremental conductance, while I /V represents instantaneous conductance.
The clearness index (K t ) is given by [66] where H ave and H o,ave are solar radiations and top-ofatmosphere radiations with the following expression. While ϕ and N are longitudes of the specified location and number of days respectively. The total PV output current (I ) can be calculated as [2] (13) where N P and N S are respectively the number of cells of PV panel array which are connected in parallel and series. I PV , and I O are photocurrent of PV cell and reverse saturation current respectively. V , and V t represent cell rated and thermal voltage while n is diode ideality factor. The intrinsic series and shunt resistances are denoted by R S and R Sh respectively.
The constraint for solar power generation is [6] The boost converter is implemented to control the performance of PV panels. the photovoltaic current (I PV ) is directly related to insolation (λ S ) as [67] where I SC , k temp , T nom , and λ nom are short circuit current (SCC), temperature coefficient related to SCC, nominal insolation and temperature respectively.

3) BATTERY ENERGY STORAGE SYSTEM (BESS)
During the intermittent renewables, the HRES reliability can be ensured by using battery storage backup. SPRE 06 415 models of the battery are used with 2.45 kWh nominal capacity, 80 % roundtrip efficiency, and 1958 kWh lifetime (throughput). The initial, replacement, and operating cost is supposed to be $ 176 /unit, $ 176/unit, and $ 8/yr respectively [68]. The minimum and maximum SOC range is 20 % and 70 % respectively with 20 % depth of discharge (DoD) and fast charging. The calculation for battery size is expressed in Eq. (16) [69].
where V , N day , η conv , η bat , and DOD are no-load voltage, the number of days without charging, converter efficiency, battery efficiency, and depth of discharge respectively. The mathematical relation for terminal voltage and SOC of the battery is expressed in Eqs. (17)-(18) [70].
The formula for the inductor design of a battery storage unit is defined in Eq. (22) [71].
The parameters for the design of battery inductor include [33] I bat which is 20 % of I bat , while I bat < I bat(max) , and D Bat = V DC /V Bat.

4) POWER CONVERTER
A dc-ac converter is used to transfer power from dc bus to ac bus with initial and replacement cost of $ 200/kW each, while O&M cost is negligible and set to zero with a lifespan of 15 years and 95 % efficiency. VOLUME 8, 2020 The relation for the power capacity (C) of the converters is shown in Eq. (22) as [2] C = (3 × L ind ) + L res (22) where L ind and L res are inductive (refrigerator, fan, pump) and resistive (LED lights, TV, mobile charger) loads respectively.

5) DIESEL GENERATOR
A DG is employed to fulfill a high-power deficit during fluctuating generation. The capital and O&M costs are considered as $ 1000/kW and $ 0.05/kW [72], with $ 0.8 fuel price per liter of diesel-based on market trends in Pakistan. A lifetime of 10000 hours and a 25 % minimum DG load ratio is considered. Depending upon output power, the fuel consumption of DG can be estimated from Eq. (23) [72]: where F DG ,α,β,C dg ,andP dg are fuel consumption rate in Ltr/hr, fuel intercept coefficient in Ltr/kWh, fuel slope in Ltr/kWh, diesel generator capacity in kW, and diesel generator output in kW respectively.

IV. OPTIMAL SIZING AND ECONOMIC ANALYSIS USING HOMER PRO
Different optimization tools compared in [73], are discussed and used in literature to design an optimally planned HRES. Out of 41 software tools mentioned in [74] viz. HYBRID 2 [75], RETScreen [76], iHOGA [77], TRNSYS [78], and IPSYS; and 6 user-friendly tools [79]; HOMER (Hybrid Optimization of Multiple Electric Renewables) software is the most frequently used and highly recommended in various research studies. HOMER is a robust techno-economic optimization tool [45] which allows adaptability to analyze and simulate techno-economic models and design optimization of HRES units [45].
Its optimization algorithms allow the designer as well as the decision-makers to estimate the feasibility in terms of economic and technical aspects with various technical selections regarding the fluctuations in technology costs as well as the availability of resources. Hence, HOMER is selected for techno-economic analysis to determine the most suitable model choice for fulfilling the load demand of the suggested location.
HOMER evaluates quickly and easily for feasible and optimal plans out of various possible solutions. The proposed optimization flow chart is shown in Fig. 13. HOMER optimizes the sizes of PV panels, the number of wind turbine units, the number of battery storage units, and the a number of converters [2]. HOMER optimization is based on the input parameters including load demand, energy resources data, economic and technical aspects of each component, design constraints, proposed management, and control strategy, emission data [2], [45] with total NPC as the objective function of optimization strategy [45]. HOMER analysis is based on one-year optimization for the evaluation of technical, environmental, and economic aspects of HRES [80]. Afterward, extrapolation of all costs works for the rest of the years throughout the project lifetime, which is established on linear depreciation and the most feasible plans to guarantee uninterruptible power supply and energy demand balance based on hourly duration [45]. After completing the testing step for all potential plans of HRES, the feasible configurations are found and graded on the basis of design objectives.
The past studies presented different evaluation criteria mentioned in [42] are used to find the feasible plan of HRES. In this work, TNPC and LCOE are used as the main objective function of the proposed HRES with mathematical relations are expressed as where, C ann,t shows the total HRES annual cost in $/yr while CRF represents capital recovery factor, j, and k are the annual interest and inflation rate in % and n is project lifetime in years, respectively and E ann,t is maximum load supplied in kWh/yr. The expression for renewables shares [107] is written as where E NR and E S are the non-renewable energy and total energy supplied to the load respectively. The power balance constraint is expressed in Eq. (29) which shows that the sum of consumed power is equal or less than the generated value [81]. where P PV and P WT are the output power of i th PV and WT unit respectively. Further, the power limits of energy generation and battery operating limits [81] is satisfied i.e.
where P max charge and P max discharge shows the maximum charge and discharge power of the battery.
The diesel generator produces harmful emissions out of which carbon dioxide is the main emission which is considered in the proposed system with the following relation in Eqs. (31)-(33) [82].

V. TECHNICAL ANALYSIS USING MATLAB A. FCS-MPC FOR RECONFIGURABLE INVERTER
Finite control set MPC is suggested for dc-ac interlinking converter control for regulation of load voltage magnitude. FCS-MPC for two-level three-leg voltage source inverters is analyzed in this paper. The presented control scheme selects the most optimal state out of all seven possible switching states to minimize the cost function. Simulation studies of MPC based control scheme show compensation of perturbation in load, source and filter parameters with output voltage is continuously tracking the reference value without sacrificing normal inverter operation. A dc-dc battery converter is applied to regulate the dc voltage. PV and wind converters are applied for extracting maximum power from PV and wind units.
Step by step control and management strategies are discussed in detail in the following section. Fig. 14 shows the concept of the MPC control strategy. the variables which are needed to handle and solve the specific problem are represented by v. The disturbance is represented with d. the control variable is expressed by c. w is the forecasted value of the processed variable. Based on the measurements of the present system (v) and disturbances/forecasted values (d), the optimizer can be simulated.
FCS-MPC scheme is applied for primary control of the interlining voltage source inverter, including power droop control, reference generator (three-phase), and inner control loop. In contrast to conventional controllers, FCS-MPC has no requirement of PI controllers to implement the inner current and outer voltage control loop or any other complex modulation steps (like PWM and SVPWM) while this online optimization scheme is quite simple and intuitive with the fast-dynamic response as compared to the traditional control schemes. The working principle of the FCS-MPVC control mechanism is discussed in the sections below.
In the first step, Clark transformation is applied to convert voltage signals from abc to α-β reference, which are the inputs for the inner loop. Measurement of RLC filter is used to generate the switching signals for interlining VSI. The measurements of RLC filters which are taken from primary control are then used to calculate instantaneous powers (P and Q) while fundamental powers are also calculated. Droop control strategy with P − V and Q − ω is implemented for regulating the ac bus voltage. The final reference signals for VSI transistors (MOSFETs) are generated by using a three-phase sinusoidal generator to regulate the voltage. Discrete-time state-space modeling is implemented with the help of RLC parameters. FCS-MPVC based algorithm is used to predict voltage values for all fourteen (14) combinations for the next sampling duration.
The continuous state-space (CSS) model is designed in the discrete-time state-space (DSS) model. MPC algorithm predicts voltage vectors for all possible combinations for the VOLUME 8, 2020  next sampling time. CSS and DSS models, input voltage vector, and objective functions are expressed in [42]. Seven switching states are tested to find the most optimal voltage vector and its corresponding signals for the IGBT switches of VSI.
The filter model (see Fig. 15) for Predictive voltage control of Fig.16 is where V c , V i, I f , R f , and L are load side voltage vector, inverter output voltage, filter current, and filter resistance, and inductance. CSS model can be expressed as To predict the voltage, DSS modeling is obtained as By for the DSS model, we obtain Where A q = e AT s , B q =  The load observer is suggested with the following cost function, Fig. 17 shows the flow chart of the applied algorithm for inverter control in islanded mode. The active power is regulated through the voltage magnitude of the inverter output, and the reactive power is controlled through the inverter frequency. Voltage amplitude (V nom ) and angular frequency (ω nom ) of the inverter output voltage is utilized to control power flow (active and reactive) through distributed energy resources (DERs) in an ac microgrid as described in expressions (40) and (41).
The coefficients (m and n) are selected based on the following relations to ensure system stability.   where P max and Q max are the maximum powers supplied by the inverter. ω and V are maximum deviations of frequency and voltage amplitude of the inverter output. The values of m and n are 0.0014 and 0.0008 respectively.

B. BUCK-BOOST BATTERY CONTROLLER
For the regulation of dc bus voltage, Fig. 18 shows the battery model while Fig. 19 shows the control of a buck-boost converter, which comprises current and voltage regulators. The SOC is controlled from being discharged below 20 % and overcharged above 70 %. The net power is the summation of generated power from PV, WT, and DG, plus the load power which can be expressed by Eq. (44).
The battery SOC management scheme along with the applied dispatch strategy is demonstrated in Fig 22. Fig 26 presents the complete Simulink model of the most optimal HRES with PV-wind-battery units.

C. BOOST CONVERTER CONTROL
Wind generator with a fixed pitch angle (i.e. zero) and variable speed is used, while the PMSG parameters and WT parameters are taken from [42]. Fig.20 shows the MPPT control for the boost converter.

VI. RESULTS AND DISCUSSIONS A. OPTIMAL DESIGN AND PERFORMANCE ANALYSIS USING HOMER PRO
The economic aspects and technical data for every HRES component including PV, WT, DG, BSS and converter based on the load profile data, real-time PV irradiance/temperature and wind speed are investigated. The following assumptions and constraints are considered for modeling and simulation of HRES shown in Fig. 21.
• Project lifetime is considered to be twenty years in order to obtain level best computation of the HRES.
• The nominal discount rating is considered to be 10 % based on trends in Pakistan with 4 % inflation.
• The maximum reliability is ensured by considering zero capacity shortage during the simulation of the system.
• A 10 % reserve is considered to compensate for the abrupt load variations and spikes. The wind energy reserve is chosen to be 15 % for compensating random speed variations.
• A carbon penalty cost of $ 20/tons is taken as a constraint.
• Energy flow between different components is controlled by applying the suggested dispatch methodology. Fig 22 shows the flow chart to generate enough power from diesel which can fulfill load demand. Battery charging and serving deferrable load is set as low priority objectives which are dependable on renewable energy generation. LF methodology is suitable for achieving maximum benefits from renewables and reducing fuel consumption i.e. reducing working hours of diesel generator. LF also saves the battery from over charging-discharging [40]. Optimized results of all feasible configuration plans of the suggested HRES based on techno-economic analysis are shown in Table 7. which are ranked based on NPC. The following are the concluding remarks about these results.
• Out of all possible configuration plans, PV-windbattery-converter has superior performance as well as VOLUME 8, 2020 • The winning configuration plan as shown in Table 6 comprises 13.4 kW PV, four (4) wind turbines (1 kW each), twenty (20) batteries (2.37 kWh), and 3.88 kW converter.
• The obtained optimal HRES plan attained the least possible NPC ($ 28,620), COE ($ 0.311/kWh), 100 % renewable penetration (RF) without any fuel consumption as well as emission. These results will definitely serve as a viable option for remote area HRES in Pakistan based on techno-economic and environmental criteria.
• The base scheme (i.e. diesel only) is the normal trend in Pakistan which is proved to be the worst-case scenario in this study with the highest NPC ($ 156,037), COE ($ 1.700/kWh). O&M cost ($ 12,666), annual fuel consumption (4,571 L/yr) and carbon emission penalty (11,965 kg/yr). Fig. 23 shows the generated electricity sharing on a monthly basis between PV and wind. Maximum PV and wind    generation is extracted to achieve the required load demand. It is observed from Fig. 23 that the PV generation is maximum in July due to highest daily radiations i.e. 6.892 kWh/m 2 followed by the month of June which has second-highest radiations i.e. 6.877 kWh/m 2 and it can also be verified from Fig. 7.
The real-time energy scheduling of generation and utilization is shown in Fig. 24 for the selected (sample) days i.e. from August 8 to August 11.
A deep investigation of Fig. 24 discloses important aspects of the proposed HRES with a winning configuration plan. It is depicted that the PV generation is slightly higher than the excess electricity generation. It shows that most of the PV generation is in excess and it can be managed for other VOLUME 8, 2020  applications like solar water heating and home cooling system. The second alternative way to use this PV excess energy is to widen the feeding remote area by including more houses which is one of the important points to keep for future research. The battery SOC remained within permissible limits above 20 % to avoid over-discharging of the battery storage units, while the maximum battery discharge power is linearly dependent on SOC pattern. The maximum load demand is observed at about 6 am (August 8) in the morning when maximum utilization of electricity is obvious due to wakeup time for cooking breakfast and preparing for children for schools and men/ladies for jobs. At this time, PV and wind are unable to fulfill the required demand and hence the battery is discharged to fulfill the energy shortage gap. At about 11 am on August 8, the PV and wind are sharing maximum energy and all this energy is fully utilized to feed the increasing demand of load with minimum charging power of the battery. The first load peak is observed at about 1800 in the evening when wind generation is maximum while PV generation is almost negligible due to the absence of the sunshine. At this stage, the energy gap is fulfilled by utilizing the battery power while keeping its charging power very low. The same trend of load variations is observed for the rest of the days i.e. August 9-11, with the highest load peaks in the evening at about 1900. Table 6 summarizes the cost results of the suggested optimal model. The initial cost of PV is three times that of wind. While battery storage units have the almost same cost as that of wind. The O&M cost of the battery is slightly more than PV, while it is four times higher than that of wind. If we compare total cost which includes capital, O&M, replacement, and salvage, then PV and battery storage have almost the same while wind has three times less cost as compared to PV and battery storage.

B. PERFORMANCE ANALYSIS OF PROPOSED ENERGY MANAGEMENT AND CONTROL SCHEME
An energy management and control schemes are implemented based on the FCS-MPC model keeping in view the fluctuating load demands and intermittent renewable generation as shown in Figure 26. The parameters of PI controllers are shown in Table 8, while Table 9 shows the ratings of PV, wind and RLC filters. Fig. 26 is analyzed to check the operation of the suggested model. Variable wind speed changing from 7.45 m/s (at 5 s) to 6.95 m/s (at 13s) is applied (see Fig. 27), while variable PV irradiance fluctuating from 980 kWh/m 2 (at 5 s) to 1200 kWh/m 2 (at 16 s) is applied. During wind speed variation at 13s, the power coefficient curve shows a slightly spike from its normal value of 0.48 which is negligible. The regulated power coefficient which exactly follows the reference value shows efficient and intelligent design and performance of the MPPT wind controller through a dc-dc converter. The maximum value of 0.48 is achieved at a tip speed ratio with a value of 8 as shown in Fig. 28 (a). Fig. 28 (b) shows the dc bus voltage which is controlled by applying a dc-dc battery converter to stabilize the voltage at a constant dc voltage of 750 V. Based on the energy   management and control strategy of Fig. 26, the primary load is continuously ON which can be verified from the load power of Fig. 29, current magnitude and load switch status as shown in Fig. 30. Battery SOC as shown in Fig. 30 remained greater than 20 % (lower limit) while the battery is continuously charging due to excess power generated from PV and wind until SOC reaches 70 % (upper limit) at 8.262 s. At this moment, the secondary load is switched ON which can also be verified from the load power of Fig. 29 (a), and three-phase load current magnitude of Fig 30 (a) while keeping the load voltage constant through FCS-MPVC as shown in Fig. 29 (b).

1) MPC CONTROL WITH SOC UPPER LIMIT USING PV-MPPT
The performance of the battery controller can be examined from Fig. 28 (b) that the minor voltage reduction at 8.262s is observed during the application of power management strategy with battery SOC. During wind power reduction at 13 s, the battery is absorbing less power to main the load requirement under low wind power while the negligible increase in dc bus voltage due to wind power reduction is observed. At 16 s, PV generation is increased from 13.4 kW to 16.6 kW due to variation in PV irradiance. The increase in dc bus voltage is comparatively higher as compared to the previous two cases when load and wind power variation is observed at 8.262 s and 13 s respectively. FCS-MPVC for interlinking converter effectively regulated the three-phase load voltage during external load disturbance. Critical load i.e. the main load is ON while non-critical loads (i.e. primary and secondary) are both OFF as shown in Figs. 33. Fig. 31 shows the power-sharing among PV, wind, load, and battery and dc bus voltage. During the start of simulation at 5 s when both PV and wind are integrated for power-sharing, ripples in PV power dc bus voltage are observed. The possible reason for these ripples is the minimum load requirement while the MPPT controller performance is not satisfactory which contributes ripples to PV power and ultimately dc bus voltage is also disturbed. The other possible reasons and its remedy may be explored in future research. At 7.595 s, battery SOC reaches a minimum set limit of 20 % and EMS enabled switch-1 to ON position to inject primary load into the system which can be seen from Fig. 31, Fig. 32, and Fig. 33 by observing load power, load current, and load switch-1 respectively.

2) MPC CONTROL WITH SOC LOWER LIMIT USING PV-MPPT
A dc bus voltage remained constant during this load disturbance. At 13 s, wind power is reduced while battery charging power is also decreased to fulfill this energy gap. Fig. 32 shows the regulated and pure sinusoidal waveform which validates the robust performance of the FCS-MPVC strategy. Secondary load remained OFF throughout the simulation because of the SOC value below its upper limit of 70 %. At 16 s, excess power generated due to the power from PV is absorbed by the battery to maintain the constant load demand and fixed wind power generation. A dc voltage is slightly increased due to high power injection from the PV generator.   by using the PI control strategy. The model is analyzed for the upper SOC limit of the battery. Fig. 34 shows a threephase load current and voltage. Power-exchange among PV, wind, battery, and load is shown in Fig. 35. At 8.869 s, battery SOC reaches 70 % (i.e. upper limit) while suggested EMS strategy switched ON the secondary load. The three-phase current and voltage waveforms have different peaks for each phase. Further, more load power ripples are observed in the case of PI control.

4) PI CONTROL WITH SOC LOWER LIMIT USING PV-MPPT
To check the performance of the suggested model with PI control, the PV-wind-battery model is simulated with the implementation of the PV-MPPT algorithm by using the PI control strategy. The model is analyzed for the lower SOC limit of the battery. Fig. 36 shows a three-phase load current

5) MPC CONTROL WITHOUT PV-MPPT
To check the operation of the suggested model without PV-MPPT, the PV-wind-battery model is simulated without implementing the PV-MPPT algorithm for performance comparison of both models by using FCS-MPVC strategy.  PV panel is connected with the rest of the HRES through the capacitor link. PV temperature is constant at 25 • (see Fig. 38) while PV irradiance is changed from 1000 kWh/m 2 (at 5 s) into 950 kWh/m 2 (at 16 s). Wind speed is set at 8.8 m/s at 5 s and decreased to 8.3 m/s at 13 s. A dc voltage is seen to be more stable as shown in Fig. 39 despite the fact that uncontrolled PV generation has variable although stable power. At 9.71 s, battery SOC crosses the upper limit of 70 % (see Fig. 40) and hence the EMS enabled switch-2 to feed the secondary load which can be observed from the load power signal of Fig. 39 and three-phase load current curve of Fig. 40. Fig. 40 (b) shows the regulated ac load voltage under the FCS-MPVC control strategy.
It is observed from Fig. 39 that the variation of wind and load has a direct impact on PV generation fluctuation.  This effect of PV fluctuation is observed at 9.71 s when the secondary load is switched ON. The second effect is observed at 13 s when wind generation capacity is reduced by reducing the wind speed, while the last impact is seen at 16 s when PV power itself is decreased by reducing the PV irradiance.

6) PI CONTROL WITHOUT PV-MPPT
To check the performance of the suggested model with PI control, the PV-wind-battery model is simulated without implementing the PV-MPPT algorithm by using the PI control strategy. Fig. 41 shows a three-phase load current and the three-phase ac voltage. Power-exchange among PV, wind, battery, and load is shown in Fig. 42   Further, the more load power ripples are observed in the case of PI control (see Fig. 42) as compared to the efficient load power profile as shown in Fig. 39 which shows better performance of the FCS-MPVC strategy.

7) POWER OSCILLATIONS
To investigate the load power ripples in specific, Fig 43 shows the comparison of load powers obtained by PI and FCS-MPVC strategies. From 5-10 s, the power ripples magnitude with PI control is 482 W and this magnitude increases to 1000 W from 10-20 s; while the power ripple magnitude in case of FCS-MPVC strategy is only 30 W which is summarized in Table 10. Increasing load demand will inject more power ripples which are evident from the higher load curve.   Fewer THD values (0.30 %) in a case when no PV-MPPT is implemented show that the PV-MPPT boost controller also contributes more harmonics. THD of 5.44 % is observed in PI-based models which are higher as compared to FCS-MPVC approaches.    It is pertinent to mention here that the PI controllers' parameters are manually tuned, and no optimization  technique is applied. Before achieving the optimal results, the effect of K p and K i on system response and stability of the system for various combinations of K p and K i are tested by trial and error method. It was observed that a proportional controller (K p ) reduces the rise time as well as a steady-state error but has no guaranty to completely eliminate the steadystate error. While an integral controller (K i ) completely eliminates the steady-state error but has worsened the transient response. So, a lot of efforts are being made to find the tradeoff between the two parameters to keep tracking the reference values of dc-link voltage and renewables' power.   comparison tool as an important metric for the cost of renewable generated power. India, Malaysia, and Pakistan have the lowest COE as compared to other countries. While COE in Turkey and Nigeria are the highest. In Pakistan, HRES with hydro is studied while wind integration is investigated in the current study. Based on the obtained results in this paper, the comparison also highlights fair agreement with other studies and presents insight into the economic feasibility of HRES in Pakistan. Although the table shows two case studies of Pakistan by suggesting HRES models, the incorporation of wind generation and the proposed HRES model for the selected area are not investigated in any research literature so far.

VII. SUGGESTED SCHEME FEATURES AND COMPARATIVE ANALYSIS WITH LITERATURE
The comparison of THD between proposed work and the literature is shown in Table 11. THD values of 0.30% for the current study are lower than that of the system presented in Ref. [38]. Wind-battery-converter system which is presented in Ref. [42] has lower THD but the integration of PV is not incorporated in that system. Table 12 shows the comparison of THD and load power ripples between PI and FCS-MPVC controllers. Low THD and power ripples of FCS-MPVC based management strategy outperform the strategy based on VOLUME 8, 2020    PI control. It is pertinent to mention here that the significant decrease in THD value during proposed FCS-MPC methodology for a particular application can be verified from the output voltage/current signals for example the difference between Fig 32 and Fig 34 etc.
The main concluding remarks of the presented study is summarized below: • The feasible configuration scheme with PV-wind-diesel-BSS-converter is achieved with the minimum NPC ($ 28,620) and COE ($ 0.311/kWh). Therefore, 100 % renewable energy generation penetration is ensured, along with no supply shortage.
• The cost of energy is observed high ($ 0.318/kWh) during diesel operation (case-2). It is therefore analyzed that DG plays a critical role in carbon emission for offgrid residential HRES.
• The NPC of the suggested system is five times less ($ 28,620) than the base scheme ($ 156,037). Hence, off-grid operation with diesel alone is not recommended, which is normally used in Pakistan.
• Zero carbon emission and fuel intake are observed for the suggested feasible winning plan while 11,965 kg/yr and 4,571 L/yr for the base case respectively.
• Promising results are observed during verification of the proposed EMS under fluctuating PV irradiance/temperature, and wind speed with abrupt load variations.
• The power quality is improved for an islanded (standalone) HRES by using FCS-MPC.
• Compared to conventional control strategy, the suggested FCS-MPC strategy is more robust towards transient response (less rise and fall time) during a fluctuating voltage supply at the dc bus.
• The reference voltage is fully tracked by the system voltage with no steady-state error while the stability of HRES operation with 0.30 % THD for load voltage is ensured which is below 5 % according to the regulations of the IEEE standard 929, 2000 and IEEE-519 standards [2].
• Better power quality is obtained with voltage regulation under steady-state and transient modes by using applied management and control scheme.
• Energy management with SOC is also comprehensively examined during fluctuating generation and variable load conditions.
• The design of PI controllers is presented by trial and error method to get maximum solar and wind power under-regulated dc bus voltage.
• An incremental conductance (IC) algorithm is applied for maximum power extraction as a conventional method reported in the literature.
• The performance of reconfigurable inverter during generation fluctuation and variable loads is validated through simulations.
• The HRES model which is proposed in this paper will be helpful for government and energy sector planners to execute suitable policies and mechanism by integrating more renewable generation for a reliable, economical, and environment-friendly HRES scenarios. This research outcome will also open the new purview for model designers, hybrid microgrid (MG) planners, and researchers to effectively and efficiently design and utilize the HRERs by keeping in view the intermittent generation profile to handle increasing, unpredictable and abrupt load demand of the society of this global world.

VIII. CONCLUSION
Standalone HRES with PV-wind-battery is proposed as the optimal and economically most viable system, as determined by techno-economic studies carried out through HOMER and MATLAB along with FCS-MPC of a reconfigurable inverter, to fulfill the residential electricity requirement of Sherani district in the Province of Baluchistan, Pakistan. Firstly, optimal sizing of HRES components and economic investigation is performed through HOMER, while simulation studies for the suggested area with practical and real data of load profile as well as weather is investigated using different costs (capital, replacement, O&M), operating life, and efficiencies of HRES components, project lifetime, meteorological data assessment, and interest rate as the input parameters; load demand, resources availability, operating reserves, allowable capacity shortage, GHG emission penalties as optimization constraints; and NPC as decision variable. Out of nine possible optimal configurations namely PV-wind-battery, PV-wind-diesel-battery, PV-battery, PV-diesel-battery, wind-diesel-battery, PV-winddiesel, PV-diesel, wind-diesel, and diesel-battery, as examined during this work, PV-wind-battery is obtained as the most feasible and economically viable configuration (i.e. winning plan) with minimum NPC ($ 28,620) and COE ($ 0.311/kWh) which shows 81.65 % reduction in cost and 100% preserving in toxic emission, while fulfilling 100 % energy demand with 67.3 % of excess energy. The proposed optimal HRES design (winning plan) comprises 13.4 kW PV, 4 kW wind, 3.88 kW converter, and 20 units of 2.37 kWh lead-acid battery. Optimal sizes of HRES components are then used to design a management and control strategy in MATLAB/Simulink with finite control set model predictive control (FCS-MPC) of reconfigurable inverter for technical analysis based on power balance between HRES elements, constant dc and ac voltages, safe operating range of battery SOC, efficient ac voltage quality, during variations of PV irradiance, wind speed, as well as load demand. The results are validated through simulations with total harmonic distortion (THD) of 0.30 % which is well below the allowable limit according to IEEE-929 and IEEE-519 standards as compared to 5.44 % THD with the conventional PI control scheme.
The presented scheme would be an assessing tool for the governments, energy sector/microgrid planners, model designers, and researchers to investigate suitable policies, mechanisms, effective and efficient design of HRESs. An increasing, unpredictable and abrupt load demand of the society can be handled by integrating more renewable generation in terms of a reliable, economical, and environment-friendly scenarios with an understanding of intermittent generation profile. The future work includes microgrid reconfiguration under inverter and rectification mode to control the voltage and frequency during the standalone mode, and power flow during the grid-connected mode.