Bulletin of Electrical Engineering and Informatics

Received Feb 24, 2022 Revised May 28, 2022 Accepted Jul 8, 2022 To date, the most common insulating oil used in oil-immersed transformers is mineral insulating (MI) oil, which is derived from petroleum. Owing to the depletion of petroleum over the years, it can be anticipated that petroleum-derived products such as MI oils will also deplete in the future. MI oils are not only non-renewable, but they are also non-biodegradable, where these oils are harmful to the environment in cases of oil spillage. Therefore, the aim of this study is to investigate the potential of mixing MI oil with natural ester insulating (NEI) oil in order to reduce the high dependency on MI oil for transformer applications. The MI and NEI oils were mixed with different mixing ratios. AC breakdown voltage test was conducted on the MI-NEI oil blends according to the ASTM D1816 standard. From the results, it is found that the following mixing ratios (30% of MI oil + 70% of NEI oil, 20% of MI oil + 80% of NEI oil) result in significant improvement in terms of the AC breakdown voltage compared with unused MI oil. The flash point and corrosivity levels of the oil blends were also examined.


INTRODUCTION
Hybrid systems renewable energies (HRES) are becoming popular in the typologies of renewable energy.A HRES is composed of two or more renewable energy sources with appropriate energy conversion technology connected together to feed power to the local load or grid, where the researchers cared to combine the wind power and PV because they are the most promising technologies for supplying load in remote and rural regions with strengthening the hybrid system with a storage system.In last years, recent researches have been developed in the area of hybrid system.Azzaoui et al. [1] have focused on increasing the costeffectiveness of control methods through the rapid adjustment of the parameters of system where was control the systems by the backstepping method and [2] to control the PV-wind hybrid system connected to grid by using sliding mode, and [3] also did control of the photoelectric generator for used in feeding of the independent wind turbine system by fuzzy logic method.
However, these models still complain of negative factors, which are: interruption of production caused by the absence of sunlight and interruption of production due to the absence of feeding the rotor through the grid when using the double feed generator.For this, the researchers study the field of achieving integrated feeding within a single system while ensuring energy storage where was done combined the control of doubly-fed induction generator (DFIG-based) wind turbine and battery energy storage system [4], improve the power quality and capacity of a battery-powered and solar-wind hybrid power system [5], and control of the photoelectric generator for used in feeding of the independent wind turbine system [3].With the multiplicity of production systems used in the one hybrid system, researchers had to develop a management system that saves energy according to priorities, as they relied on a management system based on proportional integral (PI) controller [6].
In this context, we proposed a new management system based on the sliding mode to control the power generated by the double feed generator and algorithm of the control system with the supervisor of the different production powers.So, our goal is to develop a hybrid system (wind/photovoltaic/battery group) which can ensure an uninterrupted power generation.A maximum power is extracted from the wind and PV group.The wind turbine used is based on a doubly-fed induction generator DFIG.The power generated by the wind turbine is injected directly to the grid.The PV power is used to supply the rotor of DFIG and to charge the battery group; the rest of this power is injected to the grid.The purpose of the use of the capacitor bank group is to ensure the rotor supply of the DFIG in the case of lack of photovoltaic energy and in the case of failure of the photovoltaic system.With the help of a supervisor, this system can manage the transfer of these powers to all destinations to ensure the continuity of the service.
The rest of the paper is organized in 9 sections.The second section presents the hybrid system (photovoltaic/wind/battery group).The third section is devoted to modelling the components of the wind turbine.In the fourth section, the control strategy of the DFIG is described, and the technique of sliding mode control is applied to control the active and reactive powers generated by the wind turbine.The fifth section describes the photovoltaic conversion chains, and gives the model of its different parts with the control methodology used.The sixth section presents the perturb & observe (P&O) control algorithm for tracking the maximum power delivered by the photovoltaic generator.In the seventh section, a modelling of the battery group is established.The eighth section is devoted to the strategy designed to manage the generated electrical energy by all the system.The simulation results are discussed and analysed in the ninth section.In the last section, conclusions are summarized.

THE PHOTOVOLTAIC/WIND SYSTEM
The hybrid system contains two sources of renewable energies; the first source is a 7.5 kW rated power wind turbine with DFIG directly connected to the power grid.The power generated by the DFIG is controlled by sliding mode control technique.The second source is solar energy; it is equipped with a photovoltaic conversion system; the rated power of this system is 6.8 kW.This power is used to feed the rotor of the DFIG, to charge the group of batteries and the rest is injected to the grid.The power generated by the hybrid system is managed as shown in Figure 1.Management power of renewable energy in multiple sources system to feeding the … (Kendzi Mohammed) 621

Modeling of DFIG
In the Park model the DFIG electrical equations are [11]- [13]. and: The electromagnetic torque equation:

CONTROL STRATEGY
In the two-phase reference, the active and reactive stator powers of the DFIG are defined [8], [14]: Based on the orientation of the stator flux, on d-axis: By combining ( 6), ( 7) and ( 12), we have: By substituting ( 7) in ( 6) and using (10), the rotor voltage is:

Sliding mode control
Variable structure control with sliding mode is a nonlinear robust control approach.It provides dynamic behavior with an invariance property to uncertainties; it is simple and robust to internal or external disturbances.The sliding mode control consists of three phases: Firstly: the choice of the surface sliding mode: using of the surface form proposed by Slotine [15], [16]: with: 1 , ,..., , , ,...
where e is error on the magnitude to be adjusted, λ is positive coefficient, n is the system order, X d is desired magnitude, and X is variable of state of the magnitude ordered.Secondly: ensure the convergence conditions and stability system; this is possible by the use of the Lyapunov stability criterion: Thirdly: the design of the law control which can be defined by [17]: The control law is constructed in such a way that the system reaches the sliding surface and remains there forever.Where u eq is the equivalent control and u n is the switching term.

MODELING OF THE PHOTOVOLTAIC ENERGY CONVERSION SYSTEM 5.1. Modeling a real cell
The PV cell is described by a single diode model, as shown in Figure 2. The photovoltaic panel is modeled by a current source and two resistors, the first is in series and the second is in parallel.These resistors present losses due to the connections of conductors and semiconductors.

ISSN: 2302-9285 
Management power of renewable energy in multiple sources system to feeding the … (Kendzi Mohammed) 623 Figure 2. Photovoltaic cell [18], [19] According to Kirchhoff's current law, the current equation of photovoltaic generator is [20]: (19) and: Where Io is saturation current of the unlit junction; e is the electron charge; K is Boltzmann's constant; Tc is the photovoltaic junction temperature; m is ideality factor of the junction; Vpv is voltage across the cell; Rs is series resistance; and Rp is shunt resistor.

Bus chopper booster
The boost converter is used to convert a low input voltage into a high output voltage [21], [22].It consists of a continuous source voltage, inductance L, a switch S, a diode, and two capacitors C1 and C2.The close time (αTs) and the open time ((1-α)Ts) is switched by the transistor of chopper at a constant frequency fs where: i) Ts is the switching period (Ts=1/fs) and α is the cyclic ratio of the switch (α ∈ [0, 1]).After applying Kirchhoff's laws on Figure 3, the equation systems according to Figure 4 are as (21): In the 1 st period αTs: (21) In the 2 nd period (1-α) Ts: (22) So, in the Ts period, we have: (23)   Applying the relation (23) to the system of (21) and to (22), the equations of system in Ts period can be: (24) By arranging the terms of the preceding system of ( 24), the dynamic modeling of the boost converter is given by: ( ) ( ) Management power of renewable energy in multiple sources system to feeding the … (Kendzi Mohammed) 625 From (23), the variation of the ifhac is being a linear form, so the derivative of the variable ifhac can be defined according to (23) in αTs and (1-α)Ts periods: Substituting ( 26) and ( 27) in (28), the equations of system in Ts period can be: So, the cyclic ratio chopper (α) control is given by: (30)

MODELING OF THE BATTERY
Interruptions due to day/night alternation and the absence of the sunshine are considered disadvantages.The use of a group of batteries becomes necessary for the continuity of the supply.

Model of the battery
The battery electrical model is [24]: The state of charge (SOC) of the battery is also defined by: where Cbatt is nominal capacity of battery, and Qd is the amount of charge missing compared to Cbatt.The capacity is modeled by: where Qbatt is amount of electrical charge in Coulomb, I is current through the capacity, t is operating time, C is value of the Farad capacity, and V is potential difference corresponding to a range of SOC ranging from 0% to 100%.For nb cells in series, the battery voltage is [24]: where Vbatt is voltage of battery, Ibatt is current battery, Eb is electromotive force according to the SOC, and Ri is internal resistance of an element.

Capacity model
The capacity model is obtained from the discharge current expression in 10 h (I10).It corresponds to the operating mode of the discharge capacity in 10 h (C10) [24]:

Equation of the voltage in the charge of battery
Indeed, the charge equation of the battery voltage has the same structure as (37) which shows the influence of the electromotive force and the internal resistance [25].

STRATEGY OF THE MANAGEMENT OF THE SYSTEM STUDIED BY A SUPERVISOR
To manage the generated electrical energy and guarantee the continuity of the service (no power failure), we need a supervisor, which optimizes the use of the energy produced by the manual commutation or the automatic auxiliary power supply, by opening or closing the contacts (C1-C4) as indicated in Figure 1.At first: the supervisor checks the power status of the batteries, as they must be in a fully charged mode.Secondly: the supervisor measures the value of powers in the PV system (Ppv) and rotor (Pr), and the value of Continuous tension in the system and battery storage.Thirdly: the supervisor compares the power required by the rotor and available at the batteries and the PV system level, where the appropriate supply source is determined according to the following cases: − 1 st case: if: Ppv> Pr: the supervisor must assure the supply of the DFIG rotor with the PV generator by closing the contact C1 and keeping the battery group in the charging position by closing the contact C2 and opening the contact C3 as shows in Figure 1.− 2 nd case: if: 0 <Ppv <Pr: the supervisor must assure the supply of the DFIG rotor with two sources (generator PV, battery group) for the energy compensation, by closing the contacts C1 and C3, and opening the contacts C2 and C4 as shows in Figure 1.− 3 rd case: if: Ppv=0: the supervisor must assure the supply of the DFIG rotor with the group battery (in the night), by closing the contact C3 and opening the contacts C1, C2 and C4 as shows in Figure 1. Figure 5 presents the supervisor algorithm system.Management power of renewable energy in multiple sources system to feeding the … (Kendzi Mohammed) 627

RESULTS OF SIMULATIONS
In this section, the simulation of DC voltage adaptation system, the currents and voltages rotor, the sliding mode control technique of powers DFIG generating is described in MATLAB/Simulink environment.Then, the photovoltaic system simulation is given to find the maximum power produced in PV system.Finally, to show the reliability of the proposed control strategy, the different deficiency cases is simulated.To ensure a good power supply to the inverter and the battery bank, the switching voltage (vdc) was adapted to the value of 336 V, as shown in Figure 6.At the interval t [0, 0.04], there is a transition step due to start-up.From the instant t=0.04 s, the system is stabilized at the value of vdc=336 V.
The battery bank provides a continuous voltage Vdcbatt =336 V to supply the rotor and facilitate the coupling with the system, as shown in Figure 7.The simulation of the system discovers a fast and acceptable variation of the voltage Vrabc and the current Irabc as shown in Figure 8 and Figure 9.The sliding mode control technique is used to control the DFIG stator powers (Ps and Qs).The comparison in the cases of the PI and the sliding mode control (SMC) controllers is used to show the robustness of the control system of the active and reactive powers.Najafi-Shad et al. [6] studied the control of the hybrid system by means of a PI controller, which gave satisfactory results.But the sliding mode control had previously proven its efficacy in the independent classical systems (wind system, PV system).So, we used the control by sliding mode, to proven his efficacy by the comparison between the controller PI and sliding mode where we got better results: a very fast response time, good robustness, very fast adaptation to the disturbance of the system parameters (wind, lighting, and reference parameters) as shows in Figures 10 and 11  show that the control of the active and reactive powers by SMC technique is very fast compared to the PI technique with more precision.So, the control by SMC is robust compared to the control by PI.For different states, it does can follow the scenario declared in the management strategy of the studied system.In the case of an ordinary state (maximum lighting, no fault in the photovoltaic system), the photovoltaic system generates a stable power of a value 6.8 kW (Figure 12).The first part of this power, which does not exceed 700 W at full load, is consumed by the DFIG rotor (Figure 13).The second part Prest is variable depending on the consumption of the rotor, as shown in Figure 14.It is used to charge the group of batteries in case it is discharged.The rest is injected into the network.In the case of a problem in the photovoltaic system (system failure, absence of sun, and low lighting), it is necessary to compensate for the lack of power supply Pr using the group of batteries with manual or automatic switching.) and the switching between the DC voltages (vdc, Vdcbatt) for the supply of the inverter under transient conditions.At the period t=[0 0.1] s, the power Ppv=6.75 kW is greater than the power demanded by the rotor Pr=0.39 kW (1 st case of the control strategy), so the power benefit of Prest=6.36 kW (Figure 16(a)) is injected into the network.During this period, the system is supplied by the PV group with a voltage of vdc=336 V.The battery group is not used (Vdc batt=0 V), as shown in Figure 17 (no compensation by the battery group).From t=0.1 s to t=0.2 s, the value of the power Ppv is 0.145 kW, it is lower than the power required by the rotor (Pr=0.39kW) because of lack of power due to the lack of lighting (G=70 w/m 2 ), as shown in Figures 15 and Figure 16(b).In this case, the supervisor looks for the equilibrium point, the lack of power is compensated using the battery group (Pbatt=0.245kW) with the voltage Vdc batt=336 V, as shown in Figure 17.From t=0.24 s to t=0.4 s, there is an increase of the lighting, then of the photovoltaic power Ppv=6.75 kW (first case of the control strategy).From t=0.4 s to t=0.52 s, the power genetated by photovoltaic generator is Ppv=0 kW as shown in Figure 15 (3 rd case of the control strategy) so we have a lack of power Prest=-700 kW (Figure 16(c)).It is due to a system failure or a problem at the recovery level (vdc=0V) as shown in Figure 17.The supervisor must be powered by the capacitor bank (Vdc batt=336 V).Finally, after the moment t=0.62 s (Figure 16(d)), the system returns to the normal state (1 st case of the control strategy).

CONCLUSION
In this article, has been confirmed the reliability, cost-effectiveness and fast reactions against fugitive interruptions in a hybrid electric power generation system with three generation sources (wind turbine, photovoltaic, and battery capacitors).The primary source is the wind turbine conversion chain; it is equipped with a DFIG.The active and reactive powers of DFIG are controlled using two types of control methods: conventional method (PI) and sliding mode control method.The sliding mode controller has high accuracy and robustness compared to the PI controller.The DFIG rotor supply is provided by a photovoltaic conversion chain and by the battery group.The photovoltaic conversion chain is equipped with a controller to track the maximum power point using the P&O bserved direct search method.

Figure 10 .Figure 11 .
Figure 10.Regulation of active powers by controller PI and SMC

Figure
Figure 12.PV generated power

Figure 15 .Figure 16 .Figure 17 .
Figure 15.The appearance of the variation of the powers (Ppv, Pr) in the transitory regime ) 