DOI QR코드

DOI QR Code

A Simple Power Management Scheme with Enhanced Stability for a Solar PV/Wind/Fuel Cell Fed Standalone Hybrid Power Supply using Embedded and Neural Network Controller

  • Thangavel, S. (Dept. of Electrical and Electronic Engineering, K.S.R College of Technology) ;
  • Saravanan, S. (Dept. of Electrical and Electronic Engineering, Kongu Engineering College)
  • 투고 : 2013.08.07
  • 심사 : 2014.03.21
  • 발행 : 2014.09.01

초록

This paper propose a new power conditioner topology with intelligent power management controller that integrates multiple renewable energy sources such as solar energy, wind energy and fuel cell energy with battery backup to make best use of their operating characteristics and obtain better reliability than that could be obtained by single renewable energy based power supply. The proposed embedded controller is programmed for maintaining a constant voltage at PCC, maximum power point tracking for solar PV panel and WTG and power flow control by regulating the reference currents of the controller on instantaneous basis based on the power delivered by the sources and load demand. Instantaneous variation in reference currents of the controller enhances the controller response as it accommodates the effect of continuously varying solar insolation and wind speed in the power management. The power conditioner uses a battery bank with embedded controller based online SOC estimation and battery charging system to suitably sink or source the input power based on the load demand. The simulation results of the proposed power management system for a standalone solar/WTG/fuel cell fed hybrid power supply with real time solar radiation and wind velocity data collected from solar centre, KEC for a sporadically varying load demand is presented in this paper and the results are encouraging in reliability and stability perspective.

키워드

1. Introduction

Worldwide, the power generation is majorly done using conventional energy sources and its energy reserve is very much limited, expected to disappear after few decades. The installed capacity of India as on Oct’2013 stands on 229251.74 MW where 68.25% of the energy comes from thermal power plant which emits large amount of green house gases and enhances the global warming. According to ministry of power, the expected demand in the year 2020 and 2030 is around 4.5lakh and 9lakh MW respectively. Whereas the peak power deficit as on Oct’2013 is around 6.3%. To manage the peak power deficit, the state electricity boards impose mandatory power cut for certain period of time, which mandates the use of DG to supply the power, where the cost and greenhouse gas (GHG) emission per kWh of energy generated is very high in comparison to the utility grid.

To reduce the grid dependency, GHG emissions and to enhance the reliability in continuity of supply, the renewable energy systems such as solar PV panel, wind turbine generator (WTG), fuel cell and other co-generation system based captive power generation can be installed together to operate as a hybrid standalone power supply to meet the power demand in the industry. A suitable power conditioner is very much needed for these sources to connect to the load because of non-linear I-V characteristics and its dependency on sporadically varying natural phenomena. For the proposed system, power management is an imperative function and a dedicated controller is needed to prioritize the sources for consumption and manage the power flow from sources to load for the varying load demand. In this paper, simulation and experimental analysis of a multipleinput converter [1-4] (shown in Fig. 1) with embedded controller for carrying out the functions such as MPPT control of solar PV panel and WTG, State of Charge (SOC) estimation of battery, charging / discharging of batteries based on instantaneous load demand (LD) and power management is developed. Also the simulation analysis of a neural network controller based power management system is presented.

Fig. 1.Solar/ WTG / fuel cell fed hybrid standalone power supply

 

2. Solar Panel and VSS-INR MPPT

The daily average solar energy incident over India varies from 4-6 kWh per square meter per day depending upon the location which can be used to generate power to meet the growing demand. A single diode model based PV module [5-7] (see Fig. 2) with varying insolation and temperature is developed in simulink and is shown in Fig. 3.

Fig. 2.Equivalent circuit model of PV cell

Fig. 3.Simulink model of PV cell (with varying insolation and temperature)

The I-V characteristic of a PV array is nonlinear and varies with irradiation and temperature. The insolation change affects the photon generated current and has very little effect on the open circuit voltage. Whereas the temperature variation affects the open circuit voltage and the short circuit current varies very marginally. The powervoltage characteristics of a PV panel at different irradiance levels are shown in Fig. 4.

Fig. 4.Power-voltage characteristics of PV module at different irradiance levels

In general, there is a unique point on the I-V or V-P curve, called the Maximum Power Point(MPP), at which the entire PV system (array, converter, etc.,) operates with maximum efficiency and produces its maximum output power. The location of the MPP is not known, but can be located, either through calculation models or by search algorithms. However, by incorporating maximum power point tracking (MPPT) algorithms [8], the PV system’s power transfer efficiency and reliability can be improved significantly as it can continuously maintain the operating point of the PV panel at the MPP pertaining to that irradiation and temperature.

The solar radiation data of a typical sunny day is collected from solar centre, KEC, Erode, Tamilnadu, India is shown in Fig. 5 and the data points are extracted such that one data point for every 5 minutes and hence 288 data points or a day (86400 sec) which is fed as the insolation data input to the simulated PV panel [5-7] along with the measured panel temperature. As the change in panel temperature is marginal its effect on generated voltage is negligible and the panel output current follows insolation (Eq. 2) as photon generated current is directly proportional to insolation and the voltage, current and power output of the PV panel is shown in the Fig. 6.

Fig. 5.Variation of solar insolation in a typical sunny day at solar centre, KEC

Fig. 6.Voltage, current and power output of simulated PV panel

where IPV is the photo current(A), ID the current of parallel diode(A), ISH the shunt current, q the electron charge, m the diode ideality factor, k the Boltzmann’s constant, T temperature of the panel, Rs the series resistance, RSH the shunt resistance, VPV the output voltage of panel, I the current delivered by solar panel.

MPPT system is incorporated for solar PV panel and Wind Turbine Generator (WTG) so as to ascertain the instantaneous power generated by the sources which is needed by the power management controller. Of the various Maximum Power Point Tracking (MPPT) techniques [8, 9] variable step size incremental resistance (VSS-INR) method [10] is employed because of improved response speed, accuracy and enhanced suitability for practical operating conditions due to a wider operating range. Flowchart of the control process pertaining to VSS - INR MPPT Technique is as shown in Fig. 7.

Fig. 7.Flowchart of VSS-INR MPPT technique

The controller controls the boost converter to track the MPP by repeatedly updating the operating current of the PV array via varying the duty cycle of the boost converter in a variable step size manner. The step size Sk will become very small around the MPP by which the dynamic oscillations around MPP is reduced, which enhances the steady state performance and hence the tracking accuracy. If the operating point is far from MPP, it increases the step size of the duty cycle which enables a faster tracking ability. Progressive variation of the duty cycle in steps is observed from the Fig. 8 for a change in insolation at 4sec of simulation time. The output of MPPT converter in comparison with the solar panel output is shown in the Fig. 9.

Fig. 8.Duty cycle variation with varying step size

Fig. 9.Output of MPPT converter

 

3. Wind Turbine Generator and P&O MPPT Technique

The Simulink model of WTG developed [11-17] based on asynchronous generator is shown in Fig. 10. The simulated wind turbine block uses a 2-D lookup table to compute the turbine torque output (Tm) as a function of wind speed (w_Wind) and turbine speed (w_Turb), shown in Fig. 11. As the asynchronous machine operates in generator mode, its speed is slightly above the synchronous speed (1.011 pu). According to turbine characteristics, for a 2 m/s wind speed, the turbine output torque is adjusted so as to deliver 0.5pu of power which is 200W. The torque of the wind turbine is estimated from the basic electromechanical equation i.e., the torque is power upon generator speed as in Eq. (3)

Fig. 10.Wind turbine generator

Fig. 11.Model of wind turbine

where ωm is the generator speed, Cp the co-efficient of performance, ρ the density of air, A the area swept by turbine blade, V the velocity of wind. As the system is a standalone system, a capacitor bank is connected at the output of the asynchronous generator in order to supply the reactive power required by the asynchronous generator for generation of electrical energy. The wind speed data of a day received from solar station, KEC, Perundurai, TN, India and the simulated torque output of the wind turbine for the wind data input is given in the Figs. 12, 13 respectively.

Fig. 12.Wind speed data of a typical day at KEC, Erode, TN, India

Fig. 13.Torque output of simulated wind turbine

The Perturb and Observe algorithm (P&O) based MPPT technique is developed by programming the embedded controller. As the wind velocity is continuously varying, the torque output of wind turbine, voltage and power generated in the generator and hence the impedance of the generator keeps on varying. But the load impedance in most of the cases is continuously varying and in very few cases it is constant, so as to match the source and load impedance, a power electronic converter with MPPT controller is connected parallel with the load.

For any change in the source impedance the duty cycle of the converter is varied to match the load impedance (converter and load). P&O is basically a hill climbing technique, the controller increases the duty cycle ‘α’ of the step up chopper by 0.001(Δα) for any change in the turbine torque, and if the output power of the generator increases, the controller continue to increase till the output power starts to decrease instead of increasing and vice versa if the generators output power starts decreasing. The value of ‘Δα’ is chosen as low as 0.001 so as to reduce the dynamic power oscillations around the MPP. The flowchart of P&O MPPT control is shown in the Fig. 14. The voltage, current and power output of the WTG without and with MPPT controller is shown in the Figs. 15, 16 respectively.

Fig. 14.Flow chart of MPPT control

Fig. 15.Voltage, current and power output of WTG without MPPT controller

Fig. 16.Voltage, current and power output of WTG with MPPT controller

 

4. Fuel Cell

Fuel cell (FC) have received great deal of attention recently, because of its property of zero emission of greenhouse gasses and high power density, also have unique features such as high efficiency, diversity of fuels and reusability of the exhaust heat. Proton exchange membrane FC is used in the proposed work and in order to have the output voltage around 120V at no load condition, three FC stacks with the voltage rating of 24V each with the voltage profile of 42V at ‘0’Ampere and 35V at ‘1’ Ampere, are connected in series as shown in Fig. 17.

Fig. 17.Fuel cell stack

In the proposed paper, FC is made to supply the load when the power deliverable by solar, WTG and battery is less than load demand. When FC is delivering power, the flow rate controller of the FC is adjusted to control the hydrogen supply to FC based on the command from power management controller to generate the necessary power based on demand but the issues such as cold-start problems are not considered but can be suitably managed utilizing the dynamic behavior of the battery.

 

5. Battery

Battery is used as the external leveling agent to sink / source the power based on the instantaneous load condition. The lead acid batteries are preferred for standalone applications as the maintenance and the initial costs are less. The rate of charging and the discharging of the battery are done based on the standard specifications of the battery handbook. The lead acid battery handbook illustrates that the charging current of the battery should be less than 0.1CB, where ‘CB’ is capacity of battery. For a 150Ah battery the charging current (4) should not exceed,

Also according to the battery handbook, the discharge current in tens of the seconds should not exceed (0.5- 0.7)CB and the nominal discharge is 0.1CB. Here (CB/5) is selected as the maximum discharge current. The capacity of the battery needed for delivering the power of 1.5 kW even at minimum battery voltage of 99V and the efficiency of the boost converter being 95% ( ηBoost−Conv= 0.95) shall be calculated as in (5)

CB = 159.489Ah, Hence a 150 Ah battery is selected.

The maximum battery discharge current (6) at the output of the boost converter to deliver a power of 1.5kW at the battery voltage of V Bat -min = 99V and ηBoost−Conv = 0.95 is

5.1 SOC estimation of battery

The SOC is defined as the available capacity expressed as a percentage of its rated capacity. The slip-in and slipout of the battery from conduction is also an imperative function which is performed by the power management controller and is set at 40% State of Charge (SOC) as, Depth of Discharge (DoD) to about 70-80% of its capacity shall damage the battery even if it is a deep cycle battery.

According to lead acid battery handbook, the terminal voltage is an index of determining the SOC of battery. The SOC of a 100 V battery while floating, charging and discharging is estimated from the terminal voltage and battery current are shown in Table 1 and 2. The battery current is zero while floating, positive for discharging and vice-versa for charging. The terminal voltage, magnitude and direction of the current of battery are the inputs of the embedded controller based SOC estimator shown in Fig. 18. Of the various methods, voltage based SOC measurement is best suited for online estimation of SOC [17, 18]. SOC of battery for any intermediate voltage which lie in between the standard voltages in a column pertaining to any standard charging current is estimated using linear projection of SOC against terminal voltage. For example, SOC of a 100V battery with the terminal voltage of 106V and charging current of 30A can be estimated by the Eq. (7)

Table 1.SOC Vs terminal voltage at various charging currents and at floating

Table 2.SOC Vs terminal voltage at various discharge currents

Fig. 18.Embedded controller based SOC estimator

where V is the terminal voltage of the battery at which the SOC is to be determined.

When the terminal voltage and current of the battery are not standard values of the table, i.e., the voltage and current which lie in between the values specified in table either in row or column wise, the SOC of battery at that voltage and current is ascertained by calculating the terminal voltages for the given current pertaining to standard specified values of SOC (i.e., 10%, 20%, ..etc) in using the row wise neighboring values of voltages between which the battery current falls.

For example when the SOC of battery at the charging current of 20A is required, which falls between C/5 and C/10, hence the column wise voltages (V1….V10) for the charging current of 20A is virtually created by the controller based on the Eqs. (8)-(10)….so on

where C is the capacity of battery and I the current at which the SOC is needed.

When SOC at 106V, 20A charging is to be needed, the virtually created voltages (8), (9) and (10) adjacent to 106V are substituted in the Eq. (11) which gives the present SOC of the battery.

where V is the terminal voltage of the battery at which the SOC is sought. V2, V3 the voltages adjacent to 106V.

 

6. Neural Network Controller Based Power Management System

Power management in an instantaneous basis using Neural Network (NN) controller is realized by regulating the current from sources to load. The NN controllers are programmed to maintain a constant DC voltage at the PCC and to regulate the current flow from the sources to load accommodating the instantaneous variations in solar irradiance, wind speed and load demand (Fig. 19). When the renewable energy sources are connected in parallel at PCC through DC-DC converters the magnitude of output voltage need to be same and constant (irrespective of any changes in the input or the load) for all sources in order to limit the circulating current between the sources. As the internal impedances of the sources are different, on inclusion of load the voltage at the outputs of the DC-DC converters are different, hence a NN controller is designed for each DC-DC converter to maintain a constant output voltage by adjusting the duty cycle [Vo=(1/(1-α))Vs] where ‘α’ is the duty cycle of the converter.

Fig. 19.Implementation of neural network controller for power management

Output voltage of the boost converter connected to the sources and the voltage at PCC is shown in the Fig. 20

Fig. 20.Voltage at output of boost converters at PCC

The training data of the NN designed for voltage control are error voltage (deviation from reference voltage) and the change in duty cycle ‘Δα’ necessary to reinstate the reference voltage. The error voltage is given as the input to the neural network which is programmed and trained with 40 different error voltages (shown in Table. 3) to generate ‘Δα’ i.e., the change in duty cycle which is to be added with the αk-1 to nullify the error voltage. Hence ‘α’ at instant ‘k’ is given by ‘αk = αk-1+ Δα’.

Table 3.Training data of NN controller connected to the solar PV panel fed boost converter

Current output of the boost converter connected to the sources is varied by fractionally incrementing / decrementing the duty cycle of the boost converter which is already maintaining the voltage at PCC. Input of the controller (programmed and trained with 16 different values of reference current) is the instantaneous reference current and it delivers the change in duty cycle required to reinstate the instantaneous reference current of the sources. Efficacy of the power management system in managing the power flow is shown in the Figs. 21 and 22.

Fig. 21.NN Controller duly discharges the FC along with the PV and WTG, Battery to supply the fixed LD

Fig. 22.NN controller manages the LD by suitably discharging the battery utilizing all the power generated by the PV panel and WTG

 

7. Embedded Controller Based Power Management System

Input signals of the controller are information regarding the instantaneous power delivered by sources i.e., solar PV panel, WTG, maximum deliverable power of fuel cell, present SOC of battery and Load Demand (LD) and the output signals are pulses that connect / disconnect the sources and duty cycle ‘α’ for each boost converter connected to sources. The embedded controller when turned on generates duty cycle ‘α’ to develop reference voltage ‘Vref = 156V’ at the PCC as in Eq. (12) for all the sources which enables the inverter output of 110V RMS sine wave.

where αVpcc is the duty cycle of the boost converter to develop 156V at PCC, Vref the reference voltage at PCC, VIN_S the input voltage of boost converter connected to solar panel, VIN_W the input voltage of boost converter connected to WTG, VIN_F the input voltage of boost converter connected to fuel cell.

For any deviation in the PCC voltage from 156V of the sources connected to PCC (due to variation in load demand or the variation in the input voltages) but not delivering power to the load is corrected by adjusting the duty cycle. For example when solar and wind is supplying load while fuel cell is floating, for any deviation in the output voltage of the boost converter connected to PCC, the duty cycle of fuel cell is adjusted to bring to 156V using the Eq. (13)

where αF(k+1) is the duty cycle of boost converter connected to fuel cell at k+1th instant, αFk the duty cycle at the kth instant. Duty cycle of each boost converter is separately controlled but concurrently in all boost converters by the power management controller to perform both voltage and current control to vary the power delivered by the sources. The embedded controller [19-25] is also programmed to execute the following priority order and the flow chart of control process to execute the priority in consumption is shown in the Fig. 23.

Fig. 23.Flowchart of control logic

・ The solar and wind energy should be given the first priority for consumption as it is freely available than the fuel cell energy. ・ If the LD is further more the battery is made to should be able to deliver and to absorb the power quickly. The battery should be charged when there is any excess energy in the system. ・ In order to improve the life time of the battery, the controller controls both the depth of discharge (DoD) and discharge current of the battery to the standard specified values. discharge along with the solar PV panel and WTG. ・ If the nature of the load is such that if all the above sources are not able to meet the LD, the fuel cell is made to discharge along with the remaining sources to meet the load. ・ The battery energy must always be good, i.e., the battery should be able to deliver and to absorb the power quickly. The battery should be charged when there is any excess energy in the system.. ・ In order to improve the life time of the battery, the controller controls both the depth of discharge (DoD) and discharge current of the battery to the standard specified values.

When voltage is set at 156V at PCC for all the sources, the controller compares load current with the reference current of solar PV system and WTG (Eq. 14), where the reference current is proportional to the deliverable power by the sources at reference voltage (156V) and is derived from the instantaneous power delivered by the MPPT system connected to solar PV panel and WTG as in Eq. (15)

where Ps is the power delivered by the solar PV panel, Pw the power delivered by WTG. When the ‘IL’ follows the Eq. (14), duty cycle of the boost converter connected to solar PV and WTG is adjusted suitably based on Eq. (16) to deliver the reference current. Due to continuous variation in the incident solar insolation and varying temperature on panel the ‘IS_ref’ varies, and hence the ‘ISO’ i.e., actual output current of the solar PV panel fed boost converter made equal to the ‘IS_ref’ instantly by marginally varying (incrementing / decrementing) the duty cycle of the boost converter based on the Eq. (16)

Similarly IWO i.e., actual output current of WTG is made equal to IW_ref, by the characteristic Eq. (17)

Where ‘N’ is the scaling factor that influence the faster convergence of ‘ISO’ to ‘IS_ref’. When the scaling factor is a constant of fixed magnitude, the response of the system is sluggish as it can lead to larger dynamic oscillations, hence a variable scaling factor is introduced in the proposed work. when the gap between ‘ISO’ and ‘IS_ref’ is larger, the scaling factor ‘N’ is automatically made larger which enables the faster convergence of ‘ISO’ to ‘IS_ref’, similarly when the gap between ‘ISO’ and ‘IS_ref’ is smaller, ‘N’ is smaller to enable accurate settling of actual to reference value and this process is same for all the sources. Faster convergence is mandatory for effective power management and improved stability. The magnitude of ‘N’ which is initially fixed is iteratively calculated and updated, which is based on the ratio of duty cycles of the boost converter connected to the sources delivering the LD. For example consider two sources i.e., solar PV panel and battery is supplying power to meet a fixed LD and if the power delivered by PV panel gets reduced due to source constraints while the power delivered by WTG is unchanged, the duty cycle of the battery fed boost converter is increased to compensate the power loss incurred by the solar PV panel to supply the LD.

The increment or decrement in the duty cycle of the battery fed boost converter is varied in proportion to the present power delivered or duty cycle of solar PV fed boost converter. Hence the scaling factor ‘N’ of any source is adjusted in proportion to the other source which is contributing larger power to the LD and the scaling factor NS(k) and NW(k) follows the Eqs. (18) and (19).

where

If load current follows the Eq. (14), the reference charging current of battery proportional to excess power available at PCC is estimated using Eq. (20)

when IBC_act = IBC_ref i.e., actual charging current is equal to the reference charging current of battery, duty cycle of buck converter charging the battery is maintained same by keeping the same reference voltage in duty cycle ‘α’ generation i.e., VR_pwm(BC)k+1 = VR_pwm(BC)k, in case if IBC_act ≠ IBC_ref , the reference voltage compared with the carrier wave to generate duty cycle is incremented or decremented in fixed step size to reach the reference charging current as in Eq. (21). The value of ‘ΔVR’ is chosen as 0.00001 for IBC_act >IBC_ref and 0.000001for IBC_act< IBC_ref.

When the LD is higher, battery is discharged along with solar PV panel and WTG to supply the load demand. Controller is programmed to discharge the battery only if the SOC of battery is more than 40%. Power generated by the PV panel and WTG are completely utilized and remaining power deficit to meet the LD alone is availed from battery. The discharging current reference ‘IBD_ref’ at 156V to meet the power deficit is calculated by the Eq. (22)

Duty cycle adjustment to make the actual discharging current equal to the reference discharging current i.e., ‘IBD_act = IBD_ref’ follows the Eq. (23).

The scaling factor ‘N’ gets automatically adjusted based on the Eq. (24) if ‘IS_ref > IW_ref’ and Eq. (25) if ‘IW_ref >IS_ref’ for quick convergence of ‘IBDO’ to ‘IBD_ref’ and ‘αBD(k-1)’ follows the equation similar to (18)

Similarly when solar PV, WTG, fuel cell and battery supplies the LD, the scaling factor ‘NF(k)’ of the fuel cell when ‘IS_ref > IW_ref’ is given by Eq. (26) and follows Eq. (27) if vice-versa and ‘αF(k-1)’ follows the equation similar to (19).

If the load demand is higher than the sum of power generated by solar PV panel, WTG, fuel cell and peak dischargeable power of battery, i.e., when the load current equation at 156V is as in Eq. (28), the load is disconnected from the power supply and the battery is charged with the power delivered by solar PV panel and WTG (29).

 

8. Controller Performance

The decision on inclusion of the sources for delivering the power to the load is done based on the instantaneous power delivered by the solar PV panel and WTG at the MPPT converter output, maximum power deliverable by fuel cell and present SOC of the battery. Based on the power delivered by the sources, the controller combines the sources in any of the 25 possible ways to meet the LD and is shown in the Fig. 24. Also the embedded controller is integrated with the multiple-input converter and performance of the controller is ascertained to be functioning well as programmed from the simulation output of the system for a controlled input applied at the input of various sources for varying and fixed LD.

Fig. 24.Possible combinations of sources to meet the load demand

8.1 Case I

When the instantaneous power delivered by the solar PV panel and WTG at the output of the MPPT controller is higher than the LD, the controller suitably triggers the boost converter connected to solar PV panel and WTG to harness all the power generated in the same to satisfy the load demand and charges the battery by suitably triggering the battery charger with the excess power available at the PCC which can be evidenced from the Fig. 25.

Fig. 25.Solar PV panel and WTG supplies load, and battery charges with surplus power at PCC

8.2 Case II

As the instantaneous power delivered by the solar PV panel and WTG at the output of the MPPT controller is lesser than the LD, the controller discharges the battery suitably to meet the varying load demand. Also the stability of the proposed controller is confirmed to be good for all stability issues and response of the controller for sudden load rejection is shown in the Fig. 26 where the controller shifts the converter connected to battery from discharging mode to charging mode immediately.

Fig. 26.Solar PV panel, WTG and battery supplies LD, battery charges with surplus power at PCC

8.3 Case III

In order to verify the efficacy of the controller in augmenting to the priority in utilization of sources for managing the power flow, the instantaneous power delivered by the solar PV panel and WTG at the output of the MPPT converter is made lesser than the load demand. As per the priority in consumption the controller connects and discharges the battery suitably based on the load demand when SOC of battery is more than 40%, and connects the fuel cell to PCC and controls the power produced in the fuel cell by adjusting the flow rate controller that controls the hydrogen supply to the fuel cell based on load demand when the present SOC of battery is less than 40% which is shown in the Figs. 27 and 28 respectively.

Fig. 27.Controller duly discharges the battery (as SOC is more than 40%) along with the instantaneous power delivered by solar PV panel and WTG to meet the constant LD

Fig. 28.Controller duly connects the fuel cell along with the PV and WTG to supply the fixed LD

8.4 Case IV

Performance of the controller in managing the variable load demand with peak discharging current of battery is checked by fixing the input to the controller such that the LD is time varying and larger in magnitude, SOC of battery is 100% and controller follows the priority, utilizes the power delivered by the solar PV panel and WTG completely and discharges the battery to the peak discharging current recommended by battery handbook to satisfy the load demand. As power deficit still exists the flow rate of hydrogen to the fuel cell is controlled to exactly source the power deficit to supplement the LD which can be evidenced in the Fig. 29.

Fig. 29.Controller discharges the battery with the peak discharging limit, and as the LD is still higher the controller duly discharges the fuel cell along with the PV and WTG to supply the fixed LD

 

9. Experimental Setup and Results

The experimental prototype is developed to prove the validity of the proposed power management system. The proposed system comprises of four boost converters, battery charger (buck converter), Hall Effect DC current sensors (CYHCS-ES588), 50 Hz inverter and ATmega processors for SOC calculation, voltage control and power flow control and is shown in the Fig. 30. Specification of the polycrystalline solar panel used in the proposed work for powering one of the boost converters is shown in Table 4. Power supplies are used for powering the controllers and WTG fed boost converter and 12V lead acid battery is used for fuel cell and battery powered boost converter.

Table 4.Specification of solar panel

Fig. 30.Experimental setup of the proposed system

The voltage at the output of boost converter is set at 24V and the inherent closed loop voltage controller connected to each boost converter generates duty cycle based on the input and output voltage at (k-1)th instant to maintain the PCC voltage at 24V and the duty cycle of a boost converter developed by controller is shown in Fig. 31, and a diode is connected in series at the output of each converter to limit the circulating current during transients. The structure of PCC is realized by connecting the capacitors in parallel to maintain a constant voltage. The DC voltage at the PCC and the output voltage of the 50Hz inverter are shown in the Fig. 32. The output of the battery charge controller is set at 13.5V to charge the 12V lead acid battery at its standard charging current, which is varied accordingly to vary the charging current of the battery and the output of the same is shown in the Fig. 33.

Fig. 31.Control pulses (duty cycle) of boost controller generated by the voltage controller

Fig. 32.Experimental output of the voltage at PCC and output voltage of inverter

Fig. 33.Experimental output of battery charge controller

Current at the output of boost converter connected to sources, where the solar PV panel, WTG and fuel cell that supplies the LD is shown in the Fig. 30. In the event of any load side disturbances the system reinstates to the reference current very quickly which is evidenced from the Fig.34 which enhances the stability of the overall system operation. Also the power quality issues such as voltage sag and swell is very much controlled due to instantaneous control of the reference parameters.

Fig. 34.Current delivered by the sources in meeting the load demand

 

10. Conclusions

The developed embedded and neural network controller for power management system in a hybrid standalone system is able to commend well for all the different possible combinations of input powers from the sources to meet the load demand which is evidenced from the output of controllers. Also the stability of the controller is very good in reacting for the load variations and sudden load rejections. The neural network controller shows a slightly quick response in reacting to the load and source variations as the time taken to reach the reference current is faster in neural network controller than the embedded controller, and can be seen in Figs. 21 and 22. Effort can be taken to integrate the AC supply with the other renewable energy sources at the input and an intelligent controller can be designed for efficient power management which shall be used in industries as hybrid power supply with maximum utilization of green energy than grid energy. Also the proposed system with the distribution grid at the output can be operated as micro grid with suitable controller.

참고문헌

  1. Qian, Z.; Abdel-rahman, O.; Batarse, I. An integrated four - port dc / dc converter for renewable energy applications. IEEE Trans. Power Electronics, vol.25, no.7, pp. 1877-1887, july 2010. https://doi.org/10.1109/TPEL.2010.2043119
  2. Yaow-Ming Chen; Yuan-Chuan Liu; Feng-Yu Wu. Multi-input dc / dc converter based on the multi winding transformer for renewable energy applications. IEEE Trans. Industry Applications, vol. 38, no.4, pp.1096-1104, 2002 https://doi.org/10.1109/TIA.2002.800776
  3. H. Al-Atrash, I. Batarseh, "Boost-Integrated Phase- Shift Full Bridge Converter for Three-Port Interface," Proceedings of the IEEE Power Electronics Specialist Conference 2007, pp.2313-232, 2007.
  4. Jorge L. Duarte, Marcel Hendrix, Marcelo Godoy Simoes, "Three-port bidirectional converter for hybrid fuel cell systems," IEEE Transactions on Power Electronics, vol. 22, pp. 480-487, 2007. https://doi.org/10.1109/TPEL.2006.889928
  5. Jan T. Bialasiewicz. Renewable energy systems with photovoltaic power generators: operation and modeling. IEEE Trans. Industrial Electronics, vol.55, no.7, pp. 2752-2758, July 2008. https://doi.org/10.1109/TIE.2008.920583
  6. Michael E. Ropp, Sigifredo Gonzalez, "Development of a MATLAB / simulink model of a single-phase grid-connected photovoltaic system," IEEE Trans. on Energy Conversion, vol. 24, no. 1, pp. 195-202, March 2009. https://doi.org/10.1109/TEC.2008.2003206
  7. Minwon Park, In-Keun Yu, "A novel real-time simulation technique of photovoltaic generation systems using RTDS," IEEE Trans. on Energy Conversion, vol. 19, no. 1, pp. 164-169, March 2004. https://doi.org/10.1109/TEC.2003.821837
  8. T. Esram, P. L. Chapman, "Comparison of photovoltaic array maximum power point tracking techniques," IEEE Trans. on Energy Conversion, vol. 22, no. 2, pp. 439-449, June 2007. https://doi.org/10.1109/TEC.2006.874230
  9. F. Liu, S. Duan, F. Liu, B. Liu, and Y. Kang, "A variable step size INC MPPT method for PV systems," IEEE Trans. on Industrial Electronics, vol. 55, no.7, pp. 2622-2628, July 2008. https://doi.org/10.1109/TIE.2008.920550
  10. Qiang Mei, Mingwei Shan, Liying Liu and Josep M. Guerrero, "Novel improved variable step-size incremental-resistance MPPT method for PV systems," IEEE Trans. on Industrial Electronics, vol. 58, no. 6, pp. 2427-2434, June 2011. https://doi.org/10.1109/TIE.2010.2064275
  11. A. M. Eltamaly, "Modeling of wind turbine driving permanent magnet generator with maximum power point tracking system," Journal of King Saud University - Engineering Science, vol. 19, pp. 223-237, 2007.
  12. S. M. Barakati, "Modeling and controller design of a wind energy conversion system including a matrix converter," Ph.D. Dissertation, Dept. Elect. Engg. University of Waterloo, Canada, 2008.
  13. M. H. Hansen, A. Hansen, T. J. Larsen, S. Oye, P. Sorensen, and P. Fuglsang, "Control design for a pitch-regulated, variable speed wind turbine," Riso Nat. Lab., Roskilde, Denmark, Tech. Rep. ENS1363/ 02-0017, 2005.
  14. B. Ozpineci and L. M. Tolbert, "Simulink Implementation of Induction Machine Model-a Modular Approach," Proceedings of the Electrical Machines and Drives Conference (IEMDC 2003), vol. 2, pp. 728-734, 2003.
  15. C. Liu, K. T. Chau, J. Z. Jiang, and L. Jian, "Design of a new outer-rotor permanent magnet hybrid machine for wind power generation," IEEE Trans.on Magnetics, vol. 44, no. 6, pp. 1494-1497, June 2008. https://doi.org/10.1109/TMAG.2007.916503
  16. A. Tapia, G. Tapia, J. Ostolaza, and J. Saenz, "Modeling and control of a wind turbine driven doubly fed induction generator", IEEE Trans. on Energy Conversion, vol.18, no.2, pp. 194-204, June 2003. https://doi.org/10.1109/TEC.2003.811727
  17. K. W. E. Cheng, B. P. Divakar, Hongjie Wu, Kai Ding, and Ho Fai Ho, "Battery - Management system (BMS) and SOC development for electrical vehicles," IEEE Trans. on Vehicular Technology, vol. 60, no. 1, pp. 76-88, Jan. 2011. https://doi.org/10.1109/TVT.2010.2089647
  18. S. Sato, A. kawamura, "A New Estimation Method of State of Charge using Terminal Voltage and Internal Resistance for Lead Acid Battery," Proceedings of the power conversion conference (PCC 2002), vol. 2, pp. 565-570, osaka 2002.
  19. M. N. Eskander, T. F. El-Shatter, M. T. El-Hagry, "Energy flow and management of hybrid wind / PV / fuel cell generation system," Proceedings of the IEEE 33rd Annual Power Electronics Specialists Conference (PESC 2002), vol.1, pp. 347-353, 2002.
  20. Caisheng Wang and H. Nehrir, "Power management of a stand-alone wind / photovoltaic / fuel-cell energy system", Proceedings of the IEEE Power and Energy Society General Meeting - Conversion and Delivery of Electrical Energy in the 21st Century, pp. 1, 2008.
  21. A. Hajizadeh, M. A. Golkar, "Fuzzy neural control of a hybrid fuel cell / battery distributed power generation system," IET Journal of Renewable Power Generation, vol. 3, no. 4, pp. 402-414, December 2009. https://doi.org/10.1049/iet-rpg.2008.0027
  22. S. G. Li, S. M. Sharkh, F. C. Walsh, and C. N. Zhang, "Energy and battery management of a plug-in series hybrid electric vehicle using fuzzy logic", IEEE Trans. on Vehicular Technology, vol. 60, no. 8, pp. 3571- 3585, October 2011. https://doi.org/10.1109/TVT.2011.2165571
  23. Majid Zandi, Alireza Payman, Jean-Philippe Martin, Serge Pierfederici, Bernard Davat, and Farid Meibody- Tabar, "Energy management of a fuel cell / super capacitor / battery power source for electric vehicular applications", IEEE Trans. on Vehicular Technology, vol. 60, no. 2, pp. 433-443, February 2011. https://doi.org/10.1109/TVT.2010.2091433
  24. A. Yasin, G. Napoli, M. Ferraro, A. Testa, V. Anto nucci, "Fuzzy Logic based management of a Standalone Hybrid Generator", Proceedings of the International Conference on Clean Electrical Power (ICCEP), pp. 690-696, 14-16 June 2011.
  25. M. Amirabadi, S. Farhangi, "Fuzzy control of hybrid fuel cell / battery power source in electric vehicle", IEEE Conference on Industrial Electronics and Applications, pp. 1-5, 2006.