The prosumer microgrid is made up of several kinds of energy storage facilities, loads, and some generation supplies (Diesel generator and PV). An energy supplier from a grid has a facility known as a net-metering facility for the smart prosumer, who may use it to exchange any excess energy with the grid. Fig. 1 depicts the suggested model’s conceptual model, which includes the utility grid, energy management structure, and a prosumer microgrid. The proposed EMS is focused and deployed at the prosumer end that receives load demand and pricing data, weather forecasts, the baseline condition of the energy storage, and its relevant input parameters and finds the best way to satisfy the demand of the required load using existing resources while staying within operational and design restrictions.

The controlling scheduler uses this optimum solution to allocate existing resources. The suggested EMS also has the capability of storing several essential parameters, which might be used for a variety of purposes in the future. Energy exchange data, pricing data, and smart prosumers load data are stored in market databases (DB), Real-time DB, and smart prosumer DBs. The given model is presented in the following subsections.

The proposed system’s architectural model of the University of Qatar is constructed as a mixed-integer linear optimal control problem to minimize the prosumer microgrid operational expenses while taking into consideration the lifespan of the energy storage system. The proposed model has many additional components with certain limitations and restrictions, which are mentioned in detail in the following sections.

The proposed model tries to reduce the operational costs (J) of the given prosumer microgrid, which comprises an interchange of energy costs ‘E’, the wind turbine WT’ costs, ‘DG’ generating units cost, and the energy storage deterioration (2)–(5). The total expenses for calculating the energy resources are given by Eq. (1). As shown in Eqs. (4)–(6), several factors, such as the no. of cycle used, assets investments, and entire system capacity, affects the battery life, whilst charging/discharging efficiency and power are expressed as ηchrg,ηdchrg,Ptchrg and Ptdchrg  that are signified in Eq. (5): (1)CT=J=min∑t=124(costtE+costtDG+costtBat+costtWT) (2)costtE=P(t)Gγt (3)costtDG=αTG+βP(t)DG (4)costtWT=Sc.Prated($) (5)costtBat=(Ccostn×CT×2)×(ηcp(t)chrg+p(t)dchrgηdchrg) (6)P(t)Bat=η(ch)P(t)chrg−p(t)dchrgη(dchrg)

Here, some of the factors are costs expressed as costtE,costtBat,costtWT,costtDG and it can be denoted as the cost of energy interchange, degradation of energy storage systems, wind turbine systems, and diesel generator cost at t time intervals. The electrical supply provider, Qatar Electricity & Water Co., has allocated the common TOU price connection for the institution. The energy exchange with the grid system and the power unit pricing are symbolized by P(t)G and γt at any period t. costtDG is calculated with a rated capacity of a general diesel generator as it is (T_G = 600 kW), especially fuel intercept curves with (α = 0.016 l/hr/kW), here fuel intercept curves slope has (β = 0.277 l/hr/kW) and DG power (P(t)DG) [31–33] as specified in upcoming Fig. 2. The consistent charging and discharging energy efficiency, as well as power, is denoted as η(chrg),η(dchrg),p(t)chrg, and p(t)dchrg respectively and P(t)Bat is denoted as battery power in Eq. (6).

The load equality constraints symbolize the demand-side equilibrium. The expression (7) must be met and fulfilled to attain equilibrium. However, the prosumer load profile is denoted as P(t)L.

ESS does not need to be a neglected factor in power management, since it supports to contribute the system’s power in any tragic occasion such as power system failure. Although this ESS is normally hard to charge or discharge quickly, their energy limitation has indeed been incorporated in the limitations (8)–(9). Whereas, up to (12) the battery charging state “BSOC_t” relies on the prior state represented as BSOC_(t−1), which is combined with an Eq. (13) at any time t. The extreme limitations of the battery charging status are correspondingly denoted by BSOC_min and BSOCmax and in an equation to avoid ESS overcharging and discharging issues (14). A battery state of charge (BSOC_t) at t interval (day’s end) is equal with the primary state of the battery (BSOC_o) that occurs at the beginning of a day is shown in Eq. (15). (7)P(t)G+P(t)PV+P(t)Bat+P(t)DG+P(t)WT=P(t)L (8)BSOCt−1−BSOCmax100Ces≤P(t)Bat (9)P(t)Bat≤BSOC(t−1)−BSOC(min)100Ces (10)0≤η(ch)p(t)chrg≤YtchrgP(chrg, max)Bat (11)0≤P(t)dchrgη(dchrg)≤YtdchrgP(dchrg,max)Bat (12)Ytchrg+Ytdchrg≤1∀t (13)BSOC(t)=BSOC(t−1)−100×η(dchrg)P(t)dchrgCes−100×P(t)dchrgCesη(dchrg) (14)BSOC(min)≤BSOC(t)≤BSOC(max) (15)BSOC(t)=BSOC(o)

To efficiently plan the energy contribution in an energy management system, the output power of the storage system PtBat is already incorporated into the equality requirement specified in Eq. (8). However, the ESS charging or discharging is characterized by high or low values of PtBat, respectively. The integer value variables μtchrg and μtdchrg, respectively, denote the ESS charging/discharging in an interval “t”. The specified binary variable accessible in certain formulas (11) to (13), could be “1” in comparable timings to effectively prevent the BESS charging/discharging issue. The activating mode is indicated by a value of “1” for these kinds of variables.

The energy storage’s output gradient is as follows: (16)|P(t)Bat−P(t+1)Bat|≤ΔPBat

As utilities rely on their components of load demand, they must negotiate peak load contracts through consumers regularly. Some demand that goes beyond the scope of such a contract will consequence in penalties or termination of the electricity connection. Diesel generators, likewise, cannot manage loads more than their design capacity. For the generating units and the grid-connected PV, phrases are used to address the availability of power restrictions, as shown in the Eqs. (17)–(18). (17)P(min)G≤P(t)G≤P(max)G (18)P(min)DG≤P(t)DG≤P(max)DG

The exchanged net energy with a utility grid in a day is expressed as (E(net)G) while the energy exchange between utility grid while import/export are represented as positive/negative values are characterized by P(t)G, in Eq. (19) respectively. (19)E(net)G=∑t=1t=24P(t)G×h

Solar production patterns are quite unbalanced and not regular which is dependent on an environment having high irradiance output [34]. The entire year’s data will be examined under random settings. This article makes use of an earlier constructed solar irradiance simulation [35–37]. Additionally, the curve parameters of the probability distribution function are computed. The Latin Hypercube (LHS) global sampling method may generate 365 scenarios in 24 h [38–40]. The fast-forwarding approach is normally used to reduce approximately 40 randomized random generated scenarios [41,42] for the goal of reducing the calculation or computational burden, as indicated in [42]. (20)F (I)=1σ2πe−(1−μ)22σ2 (21)P(t)PV=ηpvjβpvI

In Eq. (20), the normally distributed function or a Gaussian function [43] is utilized to develop an uncertain model for solar irradiation. While ηpv; j; I and βpv denote the solar panel’s performance (17%), irradiance patterns represented as (kW/m²), and the rooftop solar area is in (m²), respectively, whereas the standard deviations of the normal distribution are denoted as µ, while mean deviation is represented as σ. Eq. (21) depicts the output power of solar energy P(t)PV, that is based on a certain area’s solar irradiation. Fig. 3 depicts the mean and standard deviation curve of solar irradiation predictable patterns for an institutional region, but it includes the smart microgrid that has been considered. The Qatar university region’s longitude and latitude are “51.4912°E” and “25.3773°N”, respectively, corresponding to almost 5.2 kWh/m² per day [38].

Eq. (22) expresses the wind energy generation P(t) exchanged with the grid system as: (22)P(t)=(0, vw<vciPrWT×(νw−νciνr−νci)PrWT+(vw−vrvci−vr)×(PcoWT−PrWT)vr<vw<vco 0, vco<vw )

The minimal cut-in velocity needed for the wind turbine to create power is stated as (νci). The maximal cut-out velocity to generate the highest power permissible to be made is given by (vco), if such speed is surpassed, then, the only thing to avoid wind turbine damage is to turn it off. However, vr denotes the rated wind speed velocity, vw is the wind speed, PrWT is the mean rated power of wind turbines, and PcoWT is the cut-out power of the wind turbine.

The Levelized cost of energy is calculated in many aspects to undertake an efficient and equitable economic study of the systems. It is determined by dividing the entire system installation cost in dollars by the amount of energy produced (kWh). The LCOE of a specific energy storage system is mainly defined as dollars per kWh of electricity. It considers all related expenses, like installation, operating, or maintenance costs, as well as capital expenditures. It may also be thought of as the lowest price on which electrical energy generation must be sold to keep the budget in range to improve net profit throughout the lifespan of the electricity generating with an energy storage component [39]. The LCOE calculation may be written as follows: (23)LCOE=Lifecycle cost($ )Lifetime electricity production (kWh)

The general solution methodology of the MILP approach is presented as follows (see Fig. 4): Because the proposed system objective function and the system restrictions happen essentially with a linear model by a large number of additional integers variable, the MILP approach is used to handle problems like linear optimization [40]. This MILP methodology is generally used as a global optimization method for solving a variety of optimization issues in branding, planning, and optimum scheduling.

Furthermore, it is evaluated with several metaheuristic approaches that provide inferior results, while MILP produces the best ideal outcomes. Consequently, the MILP approach is widely employed in EMS optimizations [41]. The following is the MILP’s general structure in Eqs. (24) and (25): (24)minx ftx (25)t0{B. x≤bBeq.x=beqxb≤x≤yb}

In different case studies, an optimum microgrid planning and approach is shown for Qatar’s hot season as it remains the same all year long. Variations in load conditions are usually normal over the year and for convenience of study, energy patterns are taken to be the same for each season. In Qatar, the highest electricity consumption cycles are Jan and Aug, as they are the system peak periods [45]. Peak load profiles for some months like (Jan and Aug) are used for economic and financial evaluation, with worst-case scenarios considered. The best potential outcome for cost reduction is achieved by selecting the worst-case scenario. To calculate the economic advantage, PV energy can be supplied to the grid to maximize the benefits.

When an institution has a working day, the administrative loads and loads of academic blocks are intended to be higher, whereas the peak energy needs in the dormitories and resident societies are examined until the middle of the night. Table 2 shows several metrics that are related to a system, while Table 3 shows a TOU scheme’s electricity pricing data. Specifically, the daily solar data utilized is taken from [46] and data properties are demonstrated and evaluated with the help of probability distribution functions indicated in the Eq. (19). Our main goal for PDF is normally to generate the daily patterns of solar on regular basis [47,48], whereas the previously present irradiance flow predicts the PV generating output power utilizing Eq. (20).

Many case scenarios are examined here in our study case, in which energy exchange analysis and energy usage according to our load are expressed using price-based metrics. To better know the power requirement for all seasons, many scenarios have been developed. Overall load patterns of load pattern are considered also, which shows the load pattern behaviour of the campus microgrid in such a way, their residential load pattern behavior, their academic load, and general load behavior are illustrated in Figs. 5 and 6.

Case Scenario (a): In our first case scenario, the campus’s electricity needs are met entirely by the grid. The campus does not have solar, wind, ESS, or a diesel generator. The operating cost of electricity is determined to be $1630.8 using the time of use (TOU) rate. In this case scenario, LCOE is computed as 0.113$/kWh. But the findings revealed that in the first case, the day-to-day operating costs of the grid’s energy are exceptionally expensive, and this case scenario is utilized further as a study to evaluate and investigate other case scenarios.

Case Scenario (b): In the 2nd case scenario, solar is associated with an institutional smart microgrid, allowing it to both export surplus power towards the utility grid and supply the demand. Solar PV yields 8584.3 kWh, demonstrating its efficiency. The LCOE is calculated in this case by 0.052$/kWh. The net cost calculated for the electricity/day is lowered by 46.08%, to $748.5 from the base case scenario.

Case Scenario (c): The ESS is combined with the Solar and grid connection in this assumed scenario. The suggested technique is employed to estimate the net electricity cost, which is $712.9, and to scheme the charging time patterns efficiently while considering all related expenses. With the use of TOU-based tariffs and the BESS optimum scheduling process, the LCOE was computed as $0.050/kWh, also depicted in Table 3. However, the modest rise in LCOE calculation is due to BESS cost, which remains here. When compared to baseline scenario 1(a), the overall cost of power is reduced by 43.89%, which is displayed in Fig. 7.

Case Scenario (d): The smart prosumer microgrid in the fourth scenario combines a backup generator with the BESS network and Solar PV in reducing the grid peak power demand during peak hours (6:30 PM to 12:00 PM). The grid can only import power equal to 50 kW, which has limitations for the connected grid, while the power output of this diesel generator is controlled to 600 kW in the peak hours, which is displayed in the Fig. 8.

An overall electricity cost is calculated to be $590.5 each day using BESS optimum scheduling. When it is contrasted with base scenario (a), the LCOE determined in this example is 0.046$/kWh, which is 41.02% better than the previous case scenario and operates with the 8.119829 s computational time.

Case Scenario (e): Scenario 1, (e): In this case scenario, a wind turbine is connected with energy storage, rooftop solar, a diesel engine, and the utility grid. In February and March, wind speed and direction are (3–5) times greater than in the other months. The smart prosumer microgrid chose a WT with a power rating of 25 kW after examining many variables, including hub height of 36.6 m, rated wind conditions of 14 m/s, wind cut-in and cut-out velocities are 3.5 m/s and 25 m/s, respectively as 25 m/s is an optimal option. The results are presented in Fig. 9.

The LCOE for this scenario was determined to be 0.044$/kWh, which is 38.99%. The prosumer net cost electricity in this case yields 8743.5 kWh, demonstrating the best optimally scheduled efficiency. The LCOE calculated intended for rooftop solar remains 0.044$/kWh. The incorporation of the WT network with the energy system, rooftop solar, diesel engine, and grid-connected resulted in a 568.8$/day reduction in the electricity cost-benefit for the Qatar smart prosumer microgrid.

The results of existing approaches to the proposed model are summarized in Table 4.

In Table 5, ^* this does not include the expenses of ESS, Solar PV, wind system, or diesel generator, which are additional system costs. To compute this cost, the LCOE for each scenario is employed. The LCOE of photovoltaics is predicted to be 0.048 dollars per kWh [42]. The installation costs of diesel generators and energy storage are partially offset by contributing of 0.155$/kWh and 0.065$/kWh, respectively, to the given model, which also included operating and repair costs of certain energy storage systems and diesel generators in various cases. If the prosumer is mentioned inside the carbon development framework, then the prosumer will receive a receivable carbon credit (CC) [42].

The impacts of different solar PV sizes of a smart microgrid to buy electricity from the utility grid and emissions of carbon dioxide for a day are investigated. When rooftop solar is incorporated doubled, the GHG emission is also reduced by 2 times, in addition to cost savings. The graph in Fig. 10 has some colourful bars that demonstrate the multiple kinds of PV incorporation according to the proposed case scenarios and their impacts on electricity cost obtained from the grid. In all of the above cases, these calculated values can analyze the differences between their operating energy costs.

The research also shows the power ratings of system components, as well as their capital costs, service & maintenance expenses, replacement costs, return on investment, increased efficiency, and lifespan.

Rooftop Solar, BESS, converters, wind energy systems, DGs, utility grid, and other additional devices throughout the distributed generation is one of the target aspects. The impacts of various parts of the system which are presently in use are investigated. With various components joined in the system, its NPV cost is calculated. If the system included components rated at 1000 kW rooftop Solar, 550 kW diesel generator, 3500 kW capacity of the energy system, and 1400 kW converters, the net current cost was estimated to be around 12.8 million dollars as illustrated in Fig. 11.

Therefore, an ideal smart and effective solution for our microgrid has been determined as an ideal solution for the Qatar Microgrid. As, it is determined that an optimal solution for our system would be 325 kW of solar PV, 25 kW of the wind turbine, and 600 kW of diesel generators, respectively. By estimating a daily power cost of 568.8$/kWh as an ideal solution for the smart microgrid, an optimum system is implemented.

However, when compared to the other case scenarios, the savings evaluated at 38.99%, which is somehow greater than the prior case scenario’s 23% savings. It is determined that the optimum smart system with the addition of a WT, PV for the system results in larger savings, making it the best viable option for the smart prosumer microgrid.