Proposal Design of a Hybrid Solar PV-Wind-Battery Energy Storage for Standalone DC Microgrid Application

This paper presents a microgrid distributed energy resources (DERs) for a rural standalone system. It is made up of solar photovoltaic (solar PV) system, battery energy storage system (BESS), and wind turbine coupled to permanent magnet synchronous generator (WT-PMSG). The DERs are controlled by maximum power point tracking (MPPT) based proportional intergral (PI) controllers for both maximum power tracking and error feedback compensation. The MPPT uses the perturb and observe (P&O) algorithm for tracking the maximum power point of the DERs. The PI gains are tuned using the Ziegler-Nichol’s method. The developed system was built and simulated in MATLAB/Simulink under two conditions constant load, and step load changes. The controllers enabled the BESS to charge even during conditions of varying load and other environmental factors such as change of irradiance and wind speed. The reference was tracked very well by the output voltage of the DC grid. This is a useful research for electrifying the rural islanded areas, too far from the grid.


Introduction
Recent research has shown that in Tanzania, the access to electricity is limited to 35.6% of the total 56.32 million population as of year 2018 [1]. This problem of limited access to electricity can be reduced not only by grid expansions, rather, by utilizing distributed renewable energy sources (RES). These systems such as solar home systems (SHS), micro-and mini-solar plants are increasingly being used as sources of electric energy in rural areas worldwide. They are designed for use at small household demand usually in power ranging in few kilowatts. Thereby causing limitations for enterprise and other potentially larger users of electricity within rural areas. Solar photovoltaic (PV) plants need big power storage (such as batteries) to provide voltage regulation, and reduce the effects of the energy source intermittency, which adds to the cost of the installation. To reduce the cost of battery, it has been proposed to integrate several RES to form a hybrid power system [2]. Therefore, there exists a need to manage the flow of energy on these hybrid RES to ensure the reliability and availability of power supply to meet load demand. However, control of these hybrid RES systems is usually a difficult task [3]. One very complex task is to control these hybrid RES distributed energy sources (DERs) in a micro grid to maintain voltage of the micro grid within an acceptable range of ± 5% of the DC bus voltage [4]. This paper will review briefly previous proposed control architectures and point out their limitations.
The traditionally employed active-reactive (PQ) droop control has been applied in power control of a microgrid, which has success in grid-connected mode. However, the PQ droop control fails miserably when the microgrid goes into islanded mode due to poor reactive and resistive (X/R) ratio [5,6]. Another researcher [7] proposed to limit the stochastically varying DERs output by interfacing them to the DC grid using a SEPIC converter. The method works well to eliminate the effects of SEPIC voltage input (from the DERs) variation and step load changes. Further, artificial neural network (ANN) and fuzzy logic controllers (they can be summed up as artificial intelligence -AI) have been employed to satisfy load requirements from a microgrid [8]. However, the large number of connected DERs, which sometimes impose those AI controllers to fulfil conflicting requirements is fraught with limited communication. Microgrid central controller (MGCC) has been installed either on low voltage side or low voltage substation [9]. This architecture is complex, expensive, and is not feasible for implementation in DERs located in rural areas of many developing nations.
Voltage regulation on the DC bus of a microgrid has been accomplished extensively in the surveyed literature [5,[10][11][12][13][14], however, still there occurs voltage dip and swell whenever the inputs, control, and outputs (loads connected to the DC bus) change abruptly. Therefore, this paper proposes a control system for the RES to maintain the DC bus voltage at 750 V irrespective of the varying solar irradiance, DC grid load step change, and wind speed changes.

Materials and Methods
This paper considers the following energy resources constituting the microgrid DERs for a rural standalone system -solar PV plant rated 31.5 kW; wind turbine equipped with permanent synchronous generator (WT-PMSG) rated 6 kW; and battery storage rated 248 Ah (6.4 kW) respectively as shown in Figure 1. The solar PV plant gives out intermittent power from the solar irradiation at a DC voltage of about 547 V, thus the DC/DC boost converter steps up this voltage to a steady output of 750 V and feeds it to the DC bus. The WT-PMSG gives intermittent output AC voltage of 500 V, which is rectified by the AC/DC rectifier to DC voltage of 477 V, and then this voltage is stepped up to 750 V through the boost DC/DC converter. Lastly, but not the least is the battery energy storage system (BESS) at 240 V DC. The battery gets charged through the bidirectional DC/DC converter (BDC) and discharges through the same.
, , oc n T n V aV   (8) The PV I dependence on the solar irradiance G is presented by (9), where nominal irradiance is n G = 1000 W/m 2 , i K is the short circuit current/temperature coefficient,

Wind turbine modeling
The wind power w P extracted from the variable wind speed w V , with turbine blades cutting an area A is (10) as presented by Haque et al. [17]. The Betz limit p C (11) is a function of blade pitch angle  , and tip speed ratio  (12). Radius of the shaft that is coupled to the rotor of the generator is R , rotating at an angular speed w . Betz limit has a maximum value equal to 0.593 [18].
Smaller wind turbines have fixed  , therefore for this case p C is a function of  alone. For every w V , there is a different optimal  . Thus, coupling a generator with variable speeds enables maximum power extraction at different w V .

Modelling of permanent magnet synchronous generator
The permanent magnet synchronous generator (PMSG) that is coupled to the wind turbine is modeled in direct-quadrature (d-q) synchronous frame [18], by the following expressions.
Direct and quadrature axes' voltages and currents are: The stator resistance is denoted by s R , while w is the electrical angular speed. The PMSG establishes a magnetic flux linkage m  , thus the electromagnetic torque e T is computed by (15), where p is the number of pole pairs of the PMSG. Since the PMSG has a cylindrical rotor, the d L = q L , then (15) simplifies to (16).

DC/DC boost converter
The unregulated DC voltage output from the solar PV array or the rectifier of the PMSG needs to be regulated to a higher DC voltage of the DC bus as shown in Figure 1. This is accomplished by the DC/DC boost converter, whose steady state continuous current conduction mode operation is represented by (17) -(18) as found in [19][20][21]. The steady state duty cycle is denoted by  (19) - (20) respectively, following the work of [22]. The DC load is modeled by resistor R , and the boost converter is switching at a frequency of s f = 20 kHz, with an allowed voltage ripple

Maximum Power Point Tracking based Proportional and Integral Controller
This section discusses the maximum power point tracking (MPPT) control method that is used with the proportional integral (PI) controller to undertake the DC bus voltage regulation. Consider the DC/DC boost converter with a source (e.g., solar PV array), and the MPPT based PI controller shown by Figure 2. With the solar panel as shown in Table 2 Each controller will be explained separately in the following subsections.   (21) accordingly as Figure 3 shows.

Proportional integral controller
The PI controller is used as a feedback compensation control algorithm in several industrial applications worldwide. For example, a PI controller was implemented to provide compensation for DC/DC boost converter [24] and a microgrid [25]. Given a sys-

DC bus voltage regulation design for solar PV
The DC-DC Boost Converter regulates the DC bus voltage through MPPT control and PI control algorithms, discussed in the previous subsections. The MPPT control output is shown by (21), while the one from the PI is shown by (22

Solar PV with battery energy storage system
Rajasekaran and Usha Rani [27] proposed a bidirectional DC/DC (BDC) converter to interface the battery to the DC microgrid. The interface is undertaken on the DC link with the boost DC/DC converter of the solar PV as shown in Figure 5. This converter is required to protect the battery from over and under charge through suitable control algorithm. Furthermore, it has to have a wide current capacity to be capable of handling high current during low voltage operations. Similarly, this paper proposes to employ the BDC, shown by I . The BDC switch is b b Q  operates at frequency of s f or any other depending on the design is the one responsible for the boost mode. The switch that is responsible for the buck mode is c d Q  operating at s f . The battery charges in the buck mode, and discharges in the boost mode. Another research group [28] presented this converter as capable of adaptive power management and control, therefore it is suitable for this application.

Control of battery energy storage system
The charging and discharging conditions of the battery energy storage system (BESS) are tied to the state of charge (SOC), DC bus voltage, and net power ( net P ) of the microgrid DERs for the rural standalone system. The net P is calculated as (25) [18], where the solar PV power is PV P , wind power is w P and the load is L P .
In case there is an excess of power from the sources ( 0 net P  ), the excess energy is used to charge the batteries. On the other hand, if it is not sufficient to power from the sources ( 0 net P  ) the battery energy is discharging as proposed limit of SOC. This proposed battery control scheme is shown by Figure 6. If the battery SOC value is higher than 50%, the BESS discharges and supplies power to the load. The BESS charges when the SOC is less than 100%, while there is excess energy from the two sources. The battery charging and discharging control system utilizes two-loop PI structure, outer loop with PI regulating its current. The boost and buckboost converter parameters are found as per Table 2. The capacity of the BESS is 12V, 124 Ah, which are tabulated in the Appendix's Table 3 [26]. The calculated and tuned con-

DC bus voltage regulation design for wind generator
The wind turbine energy conversion system consists of the wind turbine mechanically coupled to the permanent magnet synchronous generator (WT-PMSG). The parameters of PMSG shown in Table 4 [13]. The WT-PMSG is electrically connected to the rectifier, which is further coupled to the DC/DC boost converter. The MPPT is coupled to the rectifier, and tracks the maximum power through rectifier voltage d V , and current d I

The hybrid system of solar-wind with battery energy storage system
The load demand is satisfied by the combination of solar PV, BESS, and WT-PMSG as shown in Figure 8. The WT-PMSG has the input smoothing capacitor d C , boost converter inductor w L , and its switch w Q . Figure 8. Circuit diagram of microgrid DERs for a rural standalone system

Simulations and Results Discussions
The circuit diagram of a microgrid DERs for a rural standalone system was built in MATLAB/Simulink software, after design of all system parameters and controller gains. Two case studies were simulated -(1) constant load on the DC grid, and (2) step load changes on the DC grid. Both scenarios considered constant temperature of 30 o C, solar irradiance between 800 -1000 W/m 2 as shown in Figure 9. The wind speed was varied accordingly with a mean of 12 m/s which translated to angular speed of 150 -153 rad/s, as shown in Figure 9.
4.1 Case 1 constant loading of the DC grid Figure 9 shows that the solar irradiance is 1000 w/m 2 from 0 -1 s, then it decreases linearly to 2 s. This irradiance is constant till 3.5 s, then it increases linearly to 1000 W/m 2 at 4.5 s. This irradiance is constant until the 6 s mark. The angular speed is 150 rad/s from the start till 2 s. The wind speed is increased linearly to 153 rad/s at the 3 s time. It stays constant for 1.5 s. Thereafter, it falls linearly to 150 rad/s at the 6 s mark. Figure Figure 11 shows that the load current dc I , on the grid starts from 0 at 20 A. It stays the same for one minute. Then it increases step wise to 30 A and stays the same until 2.5 s. Then it jumps to 40 A at 2.5 s until 4.0 s. Thereafter, it jumps to 50 A at the 4.0 s mark. It stays this way until the 6.0 s mark. For all these step load variations, the DC grid voltage followed the reference as shown in Figure 11. While Figure 12 shows the battery voltage being maintained at 240 V, while the battery current kept on increasing thereby charging the battery, shown by the increment of SOC in the same Figure 12. Figure 13 shows that the MPPT algorithms (for solar PV and WT-PMSG) could track the maximum power as was desired. Figure 14 shows the solar PV MPPT achieving the MPP for irradiance between 800 -1000 W/m 2 . While all these happened, the output voltage could track the reference.

Conclusions
This paper set out to design a DERs standalone microgrid that feeds DC loads. It has discussed the solar PV, BESS, and WT-PMSG and their coupling to the DC grid. The control algorithms used MPPT based PI controllers for maximum power tracking and error compensation. The MPPT employed the P&O algorithm, while the PI controller gains were tuned by Ziegler-Nichols method. Under the simulations of constant load, varying load, varying irradiance, varying wind speeds, the output voltage of the grid could track the reference fairly well.
This developed prototype could further be tested on a microgrid test bed. However, due to untenable constraints, the simulations alone are presented here. Future work will involve testing the whole DERs microgrid on the test bed and validate its performance. This is a useful research for electrification of islanded areas far from grid connectivity.  Conflicts of Interest: The authors declare no conflict of interest.